Abstract
Visual fine-tuning has garnered significant attention with the rise of pre-trained vision models. The current prevailing method, full fine-tuning, suffers from the issue of knowledge forgetting as it focuses solely on fitting the downstream training set. In this paper, we propose a novel weight rollback-based fine-tuning method called OLOR (One step Learning, One step Review). OLOR combines fine-tuning with optimizers, incorporating a weight rollback term into the weight update term at each step. This ensures consistency in the weight range of upstream and downstream models, effectively mitigating knowledge forgetting and enhancing fine-tuning performance. In addition, a layer-wise penalty is presented to employ penalty decay and the diversified decay rate to adjust the weight rollback levels of layers for adapting varying downstream tasks. Through extensive experiments on various tasks such as image classification, object detection, semantic segmentation, and instance segmentation, we demonstrate the general applicability and state-of-the-art performance of our proposed OLOR. Code is available at https://github.com/rainbow-xiao/OLOR-AAAI-2024.
Analysis and Detection of Multilingual Hate Speech Using Transformer Based Deep Learning
Abstract
Hate speech is harmful content that directly attacks or promotes hatred against members of groups or individuals based on actual or perceived aspects of identity, such as racism, religion, or sexual orientation. This can affect social life on social media platforms as hateful content shared through social media can harm both individuals and communities. As the prevalence of hate speech increases online, the demand for automated detection as an NLP task is increasing. In this work, the proposed method is using transformer-based model to detect hate speech in social media, like twitter, Facebook, WhatsApp, Instagram, etc. The proposed model is independent of languages and has been tested on Italian, English, German, Bengali. The Gold standard datasets were collected from renowned researcher Zeerak Talat, Sara Tonelli, Melanie Siegel, and Rezaul Karim. The success rate of the proposed model for hate speech detection is higher than the existing baseline and state-of-the-art models with accuracy in Bengali dataset is 89%, in English: 91%, in German dataset 91% and in Italian dataset it is 77%. The proposed algorithm shows substantial improvement to the benchmark method.
Exploring Highly Quantised Neural Networks for Intrusion Detection in Automotive CAN
Abstract
Vehicles today comprise intelligent systems like connected autonomous driving and advanced driving assistance systems (ADAS) to enhance the driving experience, which is enabled through increased connectivity to infrastructure and fusion of information from different sensing modes. However, the rising connectivity coupled with the legacy network architecture within vehicles can be exploited for launching active and passive attacks on critical vehicle systems and directly affecting the safety of passengers. Machine learning-based intrusion detection models have been shown to successfully detect multiple targeted attack vectors in recent literature, whose deployments are enabled through quantised neural networks targeting low-power platforms. Multiple models are often required to simultaneously detect multiple attack vectors, increasing the area, (resource) cost, and energy consumption. In this paper, we present a case for utilising custom-quantised MLP's (CQMLP) as a multi-class classification model, capable of detecting multiple attacks from the benign flow of controller area network (CAN) messages. The specific quantisation and neural architecture are determined through a joint design space exploration, resulting in our choice of the 2-bit precision and the n-layer MLP. Our 2-bit version is trained using Brevitas and optimised as a dataflow hardware model through the FINN toolflow from AMD/Xilinx, targeting an XCZU7EV device. We show that the 2-bit CQMLP model, when integrated as the IDS, can detect malicious attack messages (DoS, fuzzing, and spoofing attack) with a very high accuracy of 99.9%, on par with the state-of-the-art methods in the literature. Furthermore, the dataflow model can perform line rate detection at a latency of 0.11 ms from message reception while consuming 0.23 mJ/inference, making it ideally suited for integration with an ECU in critical CAN networks.
Low Complexity Turbo SIC-MMSE Detection for Orthogonal Time Frequency Space Modulation
Abstract
Recently, orthogonal time frequency space (OTFS) modulation has garnered considerable attention due to its robustness against doubly-selective wireless channels. In this paper, we propose a low-complexity iterative successive interference cancellation based minimum mean squared error (SIC-MMSE) detection algorithm for zero-padded OTFS (ZP-OTFS) modulation. In the proposed algorithm, signals are detected based on layers processed by multiple SIC-MMSE linear filters for each sub-channel, with interference on the targeted signal layer being successively canceled either by hard or soft information. To reduce the complexity of computing individual layer filter coefficients, we also propose a novel filter coefficients recycling approach in place of generating the exact form of MMSE filter weights. Moreover, we design a joint detection and decoding algorithm for ZP-OTFS to enhance error performance. Compared to the conventional SIC-MMSE detection, our proposed algorithms outperform other linear detectors, e.g., maximal ratio combining (MRC), for ZP-OTFS with up to 3 dB gain while maintaining comparable computation complexity.
The Best Ends for the Best Means: Ethical Concerns in App Reviews
Abstract
This work analyzes ethical concerns found in users' app store reviews. We performed this study because ethical concerns in mobile applications (apps) are widespread, pose severe threats to end users and society, and lack systematic analysis and methods for detection and classification. In addition, app store reviews allow practitioners to collect users' perspectives, crucial for identifying software flaws, from a geographically distributed and large-scale audience. For our analysis, we collected five million user reviews, developed a set of ethical concerns representative of user preferences, and manually labeled a sample of these reviews. We found that (1) users highly report ethical concerns about censorship, identity theft, and safety (2) user reviews with ethical concerns are longer, more popular, and lowly rated, and (3) there is high automation potential for the classification and filtering of these reviews. Our results highlight the relevance of using app store reviews for the systematic consideration of ethical concerns during software evolution.
LLM4Fuzz: Guided Fuzzing of Smart Contracts with Large Language Models
Authors: Authors: Chaofan Shou, Jing Liu, Doudou Lu, Koushik Sen
Subjects: Cryptography and Security (cs.CR); Software Engineering (cs.SE)
Abstract
As blockchain platforms grow exponentially, millions of lines of smart contract code are being deployed to manage extensive digital assets. However, vulnerabilities in this mission-critical code have led to significant exploitations and asset losses. Thorough automated security analysis of smart contracts is thus imperative. This paper introduces LLM4Fuzz to optimize automated smart contract security analysis by leveraging large language models (LLMs) to intelligently guide and prioritize fuzzing campaigns. While traditional fuzzing suffers from low efficiency in exploring the vast state space, LLM4Fuzz employs LLMs to direct fuzzers towards high-value code regions and input sequences more likely to trigger vulnerabilities. Additionally, LLM4Fuzz can leverage LLMs to guide fuzzers based on user-defined invariants, reducing blind exploration overhead. Evaluations of LLM4Fuzz on real-world DeFi projects show substantial gains in efficiency, coverage, and vulnerability detection compared to baseline fuzzing. LLM4Fuzz also uncovered five critical vulnerabilities that can lead to a loss of more than $247k.
Stability Plasticity Decoupled Fine-tuning For Few-shot end-to-end Object Detection
Authors: Authors: Yuantao Yin, Ping Yin
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Few-shot object detection(FSOD) aims to design methods to adapt object detectors efficiently with only few annotated samples. Fine-tuning has been shown to be an effective and practical approach. However, previous works often take the classical base-novel two stage fine-tuning procedure but ignore the implicit stability-plasticity contradiction among different modules. Specifically, the random re-initialized classifiers need more plasticity to adapt to novel samples. The other modules inheriting pre-trained weights demand more stability to reserve their class-agnostic knowledge. Regular fine-tuning which couples the optimization of these two parts hurts the model generalization in FSOD scenarios. In this paper, we find that this problem is prominent in the end-to-end object detector Sparse R-CNN for its multi-classifier cascaded architecture. We propose to mitigate this contradiction by a new three-stage fine-tuning procedure by introducing an addtional plasticity classifier fine-tuning(PCF) stage. We further design the multi-source ensemble(ME) technique to enhance the generalization of the model in the final fine-tuning stage. Extensive experiments verify that our method is effective in regularizing Sparse R-CNN, outperforming previous methods in the FSOD benchmark.
Abstract
While third-party libraries are extensively reused to enhance productivity during software development, they can also introduce potential security risks such as vulnerability propagation. Software composition analysis, proposed to identify reused TPLs for reducing such risks, has become an essential procedure within modern DevSecOps. As one of the mainstream SCA techniques, binary-to-source SCA identifies the third-party source projects contained in binary files via binary source code matching, which is a major challenge in reverse engineering since binary and source code exhibit substantial disparities after compilation. The existing binary-to-source SCA techniques leverage basic syntactic features that suffer from redundancy and lack robustness in the large-scale TPL dataset, leading to inevitable false positives and compromised recall. To mitigate these limitations, we introduce BinaryAI, a novel binary-to-source SCA technique with two-phase binary source code matching to capture both syntactic and semantic code features. First, BinaryAI trains a transformer-based model to produce function-level embeddings and obtain similar source functions for each binary function accordingly. Then by applying the link-time locality to facilitate function matching, BinaryAI detects the reused TPLs based on the ratio of matched source functions. Our experimental results demonstrate the superior performance of BinaryAI in terms of binary source code matching and the downstream SCA task. Specifically, our embedding model outperforms the state-of-the-art model CodeCMR, i.e., achieving 22.54% recall@1 and 0.34 MRR compared with 10.75% and 0.17 respectively. Additionally, BinaryAI outperforms all existing binary-to-source SCA tools in TPL detection, increasing the precision from 73.36% to 85.84% and recall from 59.81% to 64.98% compared with the well-recognized commercial SCA product Black Duck.
Pixel-Wise Recognition for Holistic Surgical Scene Understanding
Authors: Authors: Nicolás Ayobi, Santiago Rodríguez, Alejandra Pérez, Isabela Hernández, Nicolás Aparicio, Eugénie Dessevres, Sebastián Peña, Jessica Santander, Juan Ignacio Caicedo, Nicolás Fernández, Pablo Arbeláez
Abstract
This paper presents the Holistic and Multi-Granular Surgical Scene Understanding of Prostatectomies (GraSP) dataset, a curated benchmark that models surgical scene understanding as a hierarchy of complementary tasks with varying levels of granularity. Our approach enables a multi-level comprehension of surgical activities, encompassing long-term tasks such as surgical phases and steps recognition and short-term tasks including surgical instrument segmentation and atomic visual actions detection. To exploit our proposed benchmark, we introduce the Transformers for Actions, Phases, Steps, and Instrument Segmentation (TAPIS) model, a general architecture that combines a global video feature extractor with localized region proposals from an instrument segmentation model to tackle the multi-granularity of our benchmark. Through extensive experimentation, we demonstrate the impact of including segmentation annotations in short-term recognition tasks, highlight the varying granularity requirements of each task, and establish TAPIS's superiority over previously proposed baselines and conventional CNN-based models. Additionally, we validate the robustness of our method across multiple public benchmarks, confirming the reliability and applicability of our dataset. This work represents a significant step forward in Endoscopic Vision, offering a novel and comprehensive framework for future research towards a holistic understanding of surgical procedures.
Unfair TOS: An Automated Approach using Customized BERT
Authors: Authors: Bathini Sai Akash, Akshara Kupireddy, Lalita Bhanu Murthy
Subjects: Computation and Language (cs.CL); Computers and Society (cs.CY)
Abstract
Terms of Service (ToS) form an integral part of any agreement as it defines the legal relationship between a service provider and an end-user. Not only do they establish and delineate reciprocal rights and responsibilities, but they also provide users with information on essential aspects of contracts that pertain to the use of digital spaces. These aspects include a wide range of topics, including limitation of liability, data protection, etc. Users tend to accept the ToS without going through it before using any application or service. Such ignorance puts them in a potentially weaker situation in case any action is required. Existing methodologies for the detection or classification of unfair clauses are however obsolete and show modest performance. In this research paper, we present SOTA(State of The Art) results on unfair clause detection from ToS documents based on unprecedented Fine-tuning BERT in integration with SVC(Support Vector Classifier). The study shows proficient performance with a macro F1-score of 0.922 at unfair clause detection, and superior performance is also shown in the classification of unfair clauses by each tag. Further, a comparative analysis is performed by answering research questions on the Transformer models utilized. In order to further research and experimentation the code and results are made available on https://github.com/batking24/Unfair-TOS-An-Automated-Approach-based-on-Fine-tuning-BERT-in-conjunction-with-ML.
TreeMIL: A Multi-instance Learning Framework for Time Series Anomaly Detection with Inexact Supervision
Authors: Authors: Chen Liu, Shibo He, Haoyu Liu, Shizhong Li
Abstract
Time series anomaly detection (TSAD) plays a vital role in various domains such as healthcare, networks, and industry. Considering labels are crucial for detection but difficult to obtain, we turn to TSAD with inexact supervision: only series-level labels are provided during the training phase, while point-level anomalies are predicted during the testing phase. Previous works follow a traditional multi-instance learning (MIL) approach, which focuses on encouraging high anomaly scores at individual time steps. However, time series anomalies are not only limited to individual point anomalies, they can also be collective anomalies, typically exhibiting abnormal patterns over subsequences. To address the challenge of collective anomalies, in this paper, we propose a tree-based MIL framework (TreeMIL). We first adopt an N-ary tree structure to divide the entire series into multiple nodes, where nodes at different levels represent subsequences with different lengths. Then, the subsequence features are extracted to determine the presence of collective anomalies. Finally, we calculate point-level anomaly scores by aggregating features from nodes at different levels. Experiments conducted on seven public datasets and eight baselines demonstrate that TreeMIL achieves an average 32.3% improvement in F1- score compared to previous state-of-the-art methods. The code is available at https://github.com/fly-orange/TreeMIL.
Word-Level ASR Quality Estimation for Efficient Corpus Sampling and Post-Editing through Analyzing Attentions of a Reference-Free Metric
Authors: Authors: Golara Javadi, Kamer Ali Yuksel, Yunsu Kim, Thiago Castro Ferreira, Mohamed Al-Badrashiny
Subjects: Computation and Language (cs.CL); Sound (cs.SD); Audio and Speech Processing (eess.AS)
Abstract
In the realm of automatic speech recognition (ASR), the quest for models that not only perform with high accuracy but also offer transparency in their decision-making processes is crucial. The potential of quality estimation (QE) metrics is introduced and evaluated as a novel tool to enhance explainable artificial intelligence (XAI) in ASR systems. Through experiments and analyses, the capabilities of the NoRefER (No Reference Error Rate) metric are explored in identifying word-level errors to aid post-editors in refining ASR hypotheses. The investigation also extends to the utility of NoRefER in the corpus-building process, demonstrating its effectiveness in augmenting datasets with insightful annotations. The diagnostic aspects of NoRefER are examined, revealing its ability to provide valuable insights into model behaviors and decision patterns. This has proven beneficial for prioritizing hypotheses in post-editing workflows and fine-tuning ASR models. The findings suggest that NoRefER is not merely a tool for error detection but also a comprehensive framework for enhancing ASR systems' transparency, efficiency, and effectiveness. To ensure the reproducibility of the results, all source codes of this study are made publicly available.
DACR: Distribution-Augmented Contrastive Reconstruction for Time-Series Anomaly Detection
Authors: Authors: Lixu Wang, Shichao Xu, Xinyu Du, Qi Zhu
Abstract
Anomaly detection in time-series data is crucial for identifying faults, failures, threats, and outliers across a range of applications. Recently, deep learning techniques have been applied to this topic, but they often struggle in real-world scenarios that are complex and highly dynamic, e.g., the normal data may consist of multiple distributions, and various types of anomalies may differ from the normal data to different degrees. In this work, to tackle these challenges, we propose Distribution-Augmented Contrastive Reconstruction (DACR). DACR generates extra data disjoint from the normal data distribution to compress the normal data's representation space, and enhances the feature extractor through contrastive learning to better capture the intrinsic semantics from time-series data. Furthermore, DACR employs an attention mechanism to model the semantic dependencies among multivariate time-series features, thereby achieving more robust reconstruction for anomaly detection. Extensive experiments conducted on nine benchmark datasets in various anomaly detection scenarios demonstrate the effectiveness of DACR in achieving new state-of-the-art time-series anomaly detection.
Source Detection in Networks using the Stationary Distribution of a Markov Chain
Abstract
Nowadays, the diffusion of information through social networks is a powerful phenomenon. One common way to model diffusions in social networks is the Independent Cascade (IC) model. Given a set of infected nodes according to the IC model, a natural problem is the source detection problem, in which the goal is to identify the unique node that has started the diffusion. Maximum Likelihood Estimation (MLE) is a common approach for tackling the source detection problem, but it is computationally hard. In this work, we propose an efficient method for the source detection problem under the MLE approach, which is based on computing the stationary distribution of a Markov chain. Using simulations, we demonstrate the effectiveness of our method compared to other state-of-the-art methods from the literature, both on random and real-world networks.
ANNA: A Deep Learning Based Dataset in Heterogeneous Traffic for Autonomous Vehicles
Authors: Authors: Mahedi Kamal, Tasnim Fariha, Afrina Kabir Zinia, Md. Abu Syed, Fahim Hasan Khan, Md. Mahbubur Rahman
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Recent breakthroughs in artificial intelligence offer tremendous promise for the development of self-driving applications. Deep Neural Networks, in particular, are being utilized to support the operation of semi-autonomous cars through object identification and semantic segmentation. To assess the inadequacy of the current dataset in the context of autonomous and semi-autonomous cars, we created a new dataset named ANNA. This study discusses a custom-built dataset that includes some unidentified vehicles in the perspective of Bangladesh, which are not included in the existing dataset. A dataset validity check was performed by evaluating models using the Intersection Over Union (IOU) metric. The results demonstrated that the model trained on our custom dataset was more precise and efficient than the models trained on the KITTI or COCO dataset concerning Bangladeshi traffic. The research presented in this paper also emphasizes the importance of developing accurate and efficient object detection algorithms for the advancement of autonomous vehicles.
A Fast Parallel Approach for Neighborhood-based Link Prediction by Disregarding Large Hubs
Authors: Authors: Subhajit Sahu
Subjects: Social and Information Networks (cs.SI); Distributed, Parallel, and Cluster Computing (cs.DC)
Abstract
Link prediction can help rectify inaccuracies in community detection stemming from unaccounted-for or overlooked links within networks. Many existing works use a baseline approach, which incurs unnecessary computational costs due to its high time complexity. Further, many studies focus on smaller graphs, which can lead to misleading conclusions. The report introduces two parallel approaches, called IHub and LHub, which predict links using neighborhood-based similarity measures on large graphs. LHub is a heuristic approach, which additionally disregards large hubs - based on the idea that low-degree nodes contribute significant similarity among neighbors. On a server equipped with dual 16-core Intel Xeon Gold 6226R processors, LHub is on average 563x faster than IHub, especially on web graphs and social networks, while having similar prediction accuracy. Notably, LHub achieves a link prediction rate of 38.1M edges/s and improves performance at a rate of 1.6x for every doubling of threads.
MapChange: Enhancing Semantic Change Detection with Temporal-Invariant Historical Maps Based on Deep Triplet Network
Abstract
Semantic Change Detection (SCD) is recognized as both a crucial and challenging task in the field of image analysis. Traditional methods for SCD have predominantly relied on the comparison of image pairs. However, this approach is significantly hindered by substantial imaging differences, which arise due to variations in shooting times, atmospheric conditions, and angles. Such discrepancies lead to two primary issues: the under-detection of minor yet significant changes, and the generation of false alarms due to temporal variances. These factors often result in unchanged objects appearing markedly different in multi-temporal images. In response to these challenges, the MapChange framework has been developed. This framework introduces a novel paradigm that synergizes temporal-invariant historical map data with contemporary high-resolution images. By employing this combination, the temporal variance inherent in conventional image pair comparisons is effectively mitigated. The efficacy of the MapChange framework has been empirically validated through comprehensive testing on two public datasets. These tests have demonstrated the framework's marked superiority over existing state-of-the-art SCD methods.
Edge-Enabled Real-time Railway Track Segmentation
Authors: Authors: Chen Chenglin, Wang Fei, Yang Min, Qin Yong, Bai Yun
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Accurate and rapid railway track segmentation can assist automatic train driving and is a key step in early warning to fixed or moving obstacles on the railway track. However, certain existing algorithms tailored for track segmentation often struggle to meet the requirements of real-time and efficiency on resource-constrained edge devices. Considering this challenge, we propose an edge-enabled real-time railway track segmentation algorithm, which is optimized to be suitable for edge applications by optimizing the network structure and quantizing the model after training. Initially, Ghost convolution is introduced to reduce the complexity of the backbone, thereby achieving the extraction of key information of the interested region at a lower cost. To further reduce the model complexity and calculation, a new lightweight detection head is proposed to achieve the best balance between accuracy and efficiency. Subsequently, we introduce quantization techniques to map the model's floating-point weights and activation values into lower bit-width fixed-point representations, reducing computational demands and memory footprint, ultimately accelerating the model's inference. Finally, we draw inspiration from GPU parallel programming principles to expedite the pre-processing and post-processing stages of the algorithm by doing parallel processing. The approach is evaluated with public and challenging dataset RailSem19 and tested on Jetson Nano. Experimental results demonstrate that our enhanced algorithm achieves an accuracy level of 83.3% while achieving a real-time inference rate of 25 frames per second when the input size is 480x480, thereby effectively meeting the requirements for real-time and high-efficiency operation.
Thermal Image Calibration and Correction using Unpaired Cycle-Consistent Adversarial Networks
Abstract
Unmanned aerial vehicles (UAVs) offer a flexible and cost-effective solution for wildfire monitoring. However, their widespread deployment during wildfires has been hindered by a lack of operational guidelines and concerns about potential interference with aircraft systems. Consequently, the progress in developing deep-learning models for wildfire detection and characterization using aerial images is constrained by the limited availability, size, and quality of existing datasets. This paper introduces a solution aimed at enhancing the quality of current aerial wildfire datasets to align with advancements in camera technology. The proposed approach offers a solution to create a comprehensive, standardized large-scale image dataset. This paper presents a pipeline based on CycleGAN to enhance wildfire datasets and a novel fusion method that integrates paired RGB images as attribute conditioning in the generators of both directions, improving the accuracy of the generated images.
A Survey on African Computer Vision Datasets, Topics and Researchers
Authors: Authors: Abdul-Hakeem Omotayo, Ashery Mbilinyi, Lukman Ismaila, Houcemeddine Turki, Mahmoud Abdien, Karim Gamal, Idriss Tondji, Yvan Pimi, Naome A. Etori, Marwa M. Matar, Clifford Broni-Bediako, Abigail Oppong, Mai Gamal, Eman Ehab, Gbetondji Dovonon, Zainab Akinjobi, Daniel Ajisafe, Oluwabukola G. Adegboro, Mennatullah Siam
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Computer vision encompasses a range of tasks such as object detection, semantic segmentation, and 3D reconstruction. Despite its relevance to African communities, research in this field within Africa represents only 0.06% of top-tier publications over the past decade. This study undertakes a thorough analysis of 63,000 Scopus-indexed computer vision publications from Africa, spanning from 2012 to 2022. The aim is to provide a survey of African computer vision topics, datasets and researchers. A key aspect of our study is the identification and categorization of African Computer Vision datasets using large language models that automatically parse abstracts of these publications. We also provide a compilation of unofficial African Computer Vision datasets distributed through challenges or data hosting platforms, and provide a full taxonomy of dataset categories. Our survey also pinpoints computer vision topics trends specific to different African regions, indicating their unique focus areas. Additionally, we carried out an extensive survey to capture the views of African researchers on the current state of computer vision research in the continent and the structural barriers they believe need urgent attention. In conclusion, this study catalogs and categorizes Computer Vision datasets and topics contributed or initiated by African institutions and identifies barriers to publishing in top-tier Computer Vision venues. This survey underscores the importance of encouraging African researchers and institutions in advancing computer vision research in the continent. It also stresses on the need for research topics to be more aligned with the needs of African communities.
EK-Net:Real-time Scene Text Detection with Expand Kernel Distance
Authors: Authors: Boyuan Zhu, Fagui Liu, Xi Chen, Quan Tang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Recently, scene text detection has received significant attention due to its wide application. However, accurate detection in complex scenes of multiple scales, orientations, and curvature remains a challenge. Numerous detection methods adopt the Vatti clipping (VC) algorithm for multiple-instance training to address the issue of arbitrary-shaped text. Yet we identify several bias results from these approaches called the "shrinked kernel". Specifically, it refers to a decrease in accuracy resulting from an output that overly favors the text kernel. In this paper, we propose a new approach named Expand Kernel Network (EK-Net) with expand kernel distance to compensate for the previous deficiency, which includes three-stages regression to complete instance detection. Moreover, EK-Net not only realize the precise positioning of arbitrary-shaped text, but also achieve a trade-off between performance and speed. Evaluation results demonstrate that EK-Net achieves state-of-the-art or competitive performance compared to other advanced methods, e.g., F-measure of 85.72% at 35.42 FPS on ICDAR 2015, F-measure of 85.75% at 40.13 FPS on CTW1500.
MsSVT++: Mixed-scale Sparse Voxel Transformer with Center Voting for 3D Object Detection
Abstract
Accurate 3D object detection in large-scale outdoor scenes, characterized by considerable variations in object scales, necessitates features rich in both long-range and fine-grained information. While recent detectors have utilized window-based transformers to model long-range dependencies, they tend to overlook fine-grained details. To bridge this gap, we propose MsSVT++, an innovative Mixed-scale Sparse Voxel Transformer that simultaneously captures both types of information through a divide-and-conquer approach. This approach involves explicitly dividing attention heads into multiple groups, each responsible for attending to information within a specific range. The outputs of these groups are subsequently merged to obtain final mixed-scale features. To mitigate the computational complexity associated with applying a window-based transformer in 3D voxel space, we introduce a novel Chessboard Sampling strategy and implement voxel sampling and gathering operations sparsely using a hash map. Moreover, an important challenge stems from the observation that non-empty voxels are primarily located on the surface of objects, which impedes the accurate estimation of bounding boxes. To overcome this challenge, we introduce a Center Voting module that integrates newly voted voxels enriched with mixed-scale contextual information towards the centers of the objects, thereby improving precise object localization. Extensive experiments demonstrate that our single-stage detector, built upon the foundation of MsSVT++, consistently delivers exceptional performance across diverse datasets.
Unraveling Attacks in Machine Learning-based IoT Ecosystems: A Survey and the Open Libraries Behind Them
Authors: Authors: Chao Liu, Boxi Chen, Wei Shao, Chris Zhang, Kelvin Wong, Yi Zhang
Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI)
Abstract
The advent of the Internet of Things (IoT) has brought forth an era of unprecedented connectivity, with an estimated 80 billion smart devices expected to be in operation by the end of 2025. These devices facilitate a multitude of smart applications, enhancing the quality of life and efficiency across various domains. Machine Learning (ML) serves as a crucial technology, not only for analyzing IoT-generated data but also for diverse applications within the IoT ecosystem. For instance, ML finds utility in IoT device recognition, anomaly detection, and even in uncovering malicious activities. This paper embarks on a comprehensive exploration of the security threats arising from ML's integration into various facets of IoT, spanning various attack types including membership inference, adversarial evasion, reconstruction, property inference, model extraction, and poisoning attacks. Unlike previous studies, our work offers a holistic perspective, categorizing threats based on criteria such as adversary models, attack targets, and key security attributes (confidentiality, availability, and integrity). We delve into the underlying techniques of ML attacks in IoT environment, providing a critical evaluation of their mechanisms and impacts. Furthermore, our research thoroughly assesses 65 libraries, both author-contributed and third-party, evaluating their role in safeguarding model and data privacy. We emphasize the availability and usability of these libraries, aiming to arm the community with the necessary tools to bolster their defenses against the evolving threat landscape. Through our comprehensive review and analysis, this paper seeks to contribute to the ongoing discourse on ML-based IoT security, offering valuable insights and practical solutions to secure ML models and data in the rapidly expanding field of artificial intelligence in IoT.
Detecting Out-of-Distribution Samples via Conditional Distribution Entropy with Optimal Transport
Abstract
When deploying a trained machine learning model in the real world, it is inevitable to receive inputs from out-of-distribution (OOD) sources. For instance, in continual learning settings, it is common to encounter OOD samples due to the non-stationarity of a domain. More generally, when we have access to a set of test inputs, the existing rich line of OOD detection solutions, especially the recent promise of distance-based methods, falls short in effectively utilizing the distribution information from training samples and test inputs. In this paper, we argue that empirical probability distributions that incorporate geometric information from both training samples and test inputs can be highly beneficial for OOD detection in the presence of test inputs available. To address this, we propose to model OOD detection as a discrete optimal transport problem. Within the framework of optimal transport, we propose a novel score function known as the \emph{conditional distribution entropy} to quantify the uncertainty of a test input being an OOD sample. Our proposal inherits the merits of certain distance-based methods while eliminating the reliance on distribution assumptions, a-prior knowledge, and specific training mechanisms. Extensive experiments conducted on benchmark datasets demonstrate that our method outperforms its competitors in OOD detection.
GI-PIP: Do We Require Impractical Auxiliary Dataset for Gradient Inversion Attacks?
Abstract
Deep gradient inversion attacks expose a serious threat to Federated Learning (FL) by accurately recovering private data from shared gradients. However, the state-of-the-art heavily relies on impractical assumptions to access excessive auxiliary data, which violates the basic data partitioning principle of FL. In this paper, a novel method, Gradient Inversion Attack using Practical Image Prior (GI-PIP), is proposed under a revised threat model. GI-PIP exploits anomaly detection models to capture the underlying distribution from fewer data, while GAN-based methods consume significant more data to synthesize images. The extracted distribution is then leveraged to regulate the attack process as Anomaly Score loss. Experimental results show that GI-PIP achieves a 16.12 dB PSNR recovery using only 3.8\% data of ImageNet, while GAN-based methods necessitate over 70\%. Moreover, GI-PIP exhibits superior capability on distribution generalization compared to GAN-based methods. Our approach significantly alleviates the auxiliary data requirement on both amount and distribution in gradient inversion attacks, hence posing more substantial threat to real-world FL.
Identity-Driven Multimedia Forgery Detection via Reference Assistance
Abstract
Recent advancements in technologies, such as the 'deepfake' technique, have paved the way for the generation of various media forgeries. In response to the potential hazards of these media forgeries, many researchers engage in exploring detection methods, increasing the demand for high-quality media forgery datasets. Despite this, existing datasets have certain limitations. Firstly, most of datasets focus on the manipulation of visual modality and usually lack diversity, as only a few forgery approaches are considered. Secondly, the quality of media is often inadequate in clarity and naturalness. Meanwhile, the size of the dataset is also limited. Thirdly, while many real-world forgeries are driven by identity, the identity information of the subject in media is frequently neglected. For detection, identity information could be an essential clue to boost accuracy. Moreover, official media concerning certain identities on the Internet can serve as prior knowledge, aiding both the audience and forgery detectors in determining the true identity. Therefore, we propose an identity-driven multimedia forgery dataset, IDForge, which contains 249,138 video shots. All video shots are sourced from 324 wild videos collected of 54 celebrities from the Internet. The fake video shots involve 9 types of manipulation across visual, audio and textual modalities. Additionally, IDForge provides extra 214,438 real video shots as a reference set for the 54 celebrities. Correspondingly, we design an effective multimedia detection network, Reference-assisted Multimodal Forgery Detection Network (R-MFDN). Through extensive experiments on the proposed dataset, we demonstrate the effectiveness of R-MFDN on the multimedia detection task.
Concealed Object Segmentation with Hierarchical Coherence Modeling
Authors: Authors: Fengyang Xiao, Pan Zhang, Chunming He, Runze Hu, Yutao Liu
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Concealed object segmentation (COS) is a challenging task that involves localizing and segmenting those concealed objects that are visually blended with their surrounding environments. Despite achieving remarkable success, existing COS segmenters still struggle to achieve complete segmentation results in extremely concealed scenarios. In this paper, we propose a Hierarchical Coherence Modeling (HCM) segmenter for COS, aiming to address this incomplete segmentation limitation. In specific, HCM promotes feature coherence by leveraging the intra-stage coherence and cross-stage coherence modules, exploring feature correlations at both the single-stage and contextual levels. Additionally, we introduce the reversible re-calibration decoder to detect previously undetected parts in low-confidence regions, resulting in further enhancing segmentation performance. Extensive experiments conducted on three COS tasks, including camouflaged object detection, polyp image segmentation, and transparent object detection, demonstrate the promising results achieved by the proposed HCM segmenter.
Deep Learning for Computer Vision based Activity Recognition and Fall Detection of the Elderly: a Systematic Review
Authors: Authors: F. Xavier Gaya-Morey, Cristina Manresa-Yee, Jose M. Buades-Rubio
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
As the percentage of elderly people in developed countries increases worldwide, the healthcare of this collective is a worrying matter, especially if it includes the preservation of their autonomy. In this direction, many studies are being published on Ambient Assisted Living (AAL) systems, which help to reduce the preoccupations raised by the independent living of the elderly. In this study, a systematic review of the literature is presented on fall detection and Human Activity Recognition (HAR) for the elderly, as the two main tasks to solve to guarantee the safety of elderly people living alone. To address the current tendency to perform these two tasks, the review focuses on the use of Deep Learning (DL) based approaches on computer vision data. In addition, different collections of data like DL models, datasets or hardware (e.g. depth or thermal cameras) are gathered from the reviewed studies and provided for reference in future studies. Strengths and weaknesses of existing approaches are also discussed and, based on them, our recommendations for future works are provided.
Symbrain: A large-scale dataset of MRI images for neonatal brain symmetry analysis
Authors: Authors: Arnaud Gucciardi, Safouane El Ghazouali, Francesca Venturini, Vida Groznik, Umberto Michelucci
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
This paper presents an annotated dataset of brain MRI images designed to advance the field of brain symmetry study. Magnetic resonance imaging (MRI) has gained interest in analyzing brain symmetry in neonatal infants, and challenges remain due to the vast size differences between fetal and adult brains. Classification methods for brain structural MRI use scales and visual cues to assess hemisphere symmetry, which can help diagnose neonatal patients by comparing hemispheres and anatomical regions of interest in the brain. Using the Developing Human Connectome Project dataset, this work presents a dataset comprising cerebral images extracted as slices across selected portions of interest for clinical evaluation . All the extracted images are annotated with the brain's midline. All the extracted images are annotated with the brain's midline. From the assumption that a decrease in symmetry is directly related to possible clinical pathologies, the dataset can contribute to a more precise diagnosis because it can be used to train deep learning model application in neonatal cerebral MRI anomaly detection from postnatal infant scans thanks to computer vision. Such models learn to identify and classify anomalies by identifying potential asymmetrical patterns in medical MRI images. Furthermore, this dataset can contribute to the research and development of methods using the relative symmetry of the two brain hemispheres for crucial diagnosis and treatment planning.
MInD: Improving Multimodal Sentiment Analysis via Multimodal Information Disentanglement
Authors: Authors: Weichen Dai, Xingyu Li, Pengbo Hu, Zeyu Wang, Ji Qi, Jianlin Peng, Yi Zhou
Abstract
Learning effective joint representations has been a central task in multimodal sentiment analysis. Previous methods focus on leveraging the correlations between different modalities and enhancing performance through sophisticated fusion techniques. However, challenges still exist due to the inherent heterogeneity of distinct modalities, which may lead to distributional gap, impeding the full exploitation of inter-modal information and resulting in redundancy and impurity in the information extracted from features. To address this problem, we introduce the Multimodal Information Disentanglement (MInD) approach. MInD decomposes the multimodal inputs into a modality-invariant component, a modality-specific component, and a remnant noise component for each modality through a shared encoder and multiple private encoders. The shared encoder aims to explore the shared information and commonality across modalities, while the private encoders are deployed to capture the distinctive information and characteristic features. These representations thus furnish a comprehensive perspective of the multimodal data, facilitating the fusion process instrumental for subsequent prediction tasks. Furthermore, MInD improves the learned representations by explicitly modeling the task-irrelevant noise in an adversarial manner. Experimental evaluations conducted on benchmark datasets, including CMU-MOSI, CMU-MOSEI, and UR-Funny, demonstrate MInD's superior performance over existing state-of-the-art methods in both multimodal emotion recognition and multimodal humor detection tasks.
Rethinking Centered Kernel Alignment in Knowledge Distillation
Abstract
Knowledge distillation has emerged as a highly effective method for bridging the representation discrepancy between large-scale models and lightweight models. Prevalent approaches involve leveraging appropriate metrics to minimize the divergence or distance between the knowledge extracted from the teacher model and the knowledge learned by the student model. Centered Kernel Alignment (CKA) is widely used to measure representation similarity and has been applied in several knowledge distillation methods. However, these methods are complex and fail to uncover the essence of CKA, thus not answering the question of how to use CKA to achieve simple and effective distillation properly. This paper first provides a theoretical perspective to illustrate the effectiveness of CKA, which decouples CKA to the upper bound of Maximum Mean Discrepancy~(MMD) and a constant term. Drawing from this, we propose a novel Relation-Centered Kernel Alignment~(RCKA) framework, which practically establishes a connection between CKA and MMD. Furthermore, we dynamically customize the application of CKA based on the characteristics of each task, with less computational source yet comparable performance than the previous methods. The extensive experiments on the CIFAR-100, ImageNet-1k, and MS-COCO demonstrate that our method achieves state-of-the-art performance on almost all teacher-student pairs for image classification and object detection, validating the effectiveness of our approaches.
End-to-end Multi-Instance Robotic Reaching from Monocular Vision
Authors: Authors: Zheyu Zhuang, Xin Yu, Robert Mahony
Abstract
Multi-instance scenes are especially challenging for end-to-end visuomotor (image-to-control) learning algorithms. "Pipeline" visual servo control algorithms use separate detection, selection and servo stages, allowing algorithms to focus on a single object instance during servo control. End-to-end systems do not have separate detection and selection stages and need to address the visual ambiguities introduced by the presence of arbitrary number of visually identical or similar objects during servo control. However, end-to-end schemes avoid embedding errors from detection and selection stages in the servo control behaviour, are more dynamically robust to changing scenes, and are algorithmically simpler. In this paper, we present a real-time end-to-end visuomotor learning algorithm for multi-instance reaching. The proposed algorithm uses a monocular RGB image and the manipulator's joint angles as the input to a light-weight fully-convolutional network (FCN) to generate control candidates. A key innovation of the proposed method is identifying the optimal control candidate by regressing a control-Lyapunov function (cLf) value. The multi-instance capability emerges naturally from the stability analysis associated with the cLf formulation. We demonstrate the proposed algorithm effectively reaching and grasping objects from different categories on a table-top amid other instances and distractors from an over-the-shoulder monocular RGB camera. The network is able to run up to approximately 160 fps during inference on one GTX 1080 Ti GPU.
A Review of Physics-Informed Machine Learning Methods with Applications to Condition Monitoring and Anomaly Detection
Authors: Authors: Yuandi Wu, Brett Sicard, Stephen Andrew Gadsden
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Systems and Control (eess.SY)
Abstract
This study presents a comprehensive overview of PIML techniques in the context of condition monitoring. The central concept driving PIML is the incorporation of known physical laws and constraints into machine learning algorithms, enabling them to learn from available data while remaining consistent with physical principles. Through fusing domain knowledge with data-driven learning, PIML methods offer enhanced accuracy and interpretability in comparison to purely data-driven approaches. In this comprehensive survey, detailed examinations are performed with regard to the methodology by which known physical principles are integrated within machine learning frameworks, as well as their suitability for specific tasks within condition monitoring. Incorporation of physical knowledge into the ML model may be realized in a variety of methods, with each having its unique advantages and drawbacks. The distinct advantages and limitations of each methodology for the integration of physics within data-driven models are detailed, considering factors such as computational efficiency, model interpretability, and generalizability to different systems in condition monitoring and fault detection. Several case studies and works of literature utilizing this emerging concept are presented to demonstrate the efficacy of PIML in condition monitoring applications. From the literature reviewed, the versatility and potential of PIML in condition monitoring may be demonstrated. Novel PIML methods offer an innovative solution for addressing the complexities of condition monitoring and associated challenges. This comprehensive survey helps form the foundation for future work in the field. As the technology continues to advance, PIML is expected to play a crucial role in enhancing maintenance strategies, system reliability, and overall operational efficiency in engineering systems.
Detect-Order-Construct: A Tree Construction based Approach for Hierarchical Document Structure Analysis
Authors: Authors: Jiawei Wang, Kai Hu, Zhuoyao Zhong, Lei Sun, Qiang Huo
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Document structure analysis (aka document layout analysis) is crucial for understanding the physical layout and logical structure of documents, with applications in information retrieval, document summarization, knowledge extraction, etc. In this paper, we concentrate on Hierarchical Document Structure Analysis (HDSA) to explore hierarchical relationships within structured documents created using authoring software employing hierarchical schemas, such as LaTeX, Microsoft Word, and HTML. To comprehensively analyze hierarchical document structures, we propose a tree construction based approach that addresses multiple subtasks concurrently, including page object detection (Detect), reading order prediction of identified objects (Order), and the construction of intended hierarchical structure (Construct). We present an effective end-to-end solution based on this framework to demonstrate its performance. To assess our approach, we develop a comprehensive benchmark called Comp-HRDoc, which evaluates the above subtasks simultaneously. Our end-to-end system achieves state-of-the-art performance on two large-scale document layout analysis datasets (PubLayNet and DocLayNet), a high-quality hierarchical document structure reconstruction dataset (HRDoc), and our Comp-HRDoc benchmark. The Comp-HRDoc benchmark will be released to facilitate further research in this field.
Large receptive field strategy and important feature extraction strategy in 3D object detection
Authors: Authors: Leichao Cui, Xiuxian Li, Min Meng
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
The enhancement of 3D object detection is pivotal for precise environmental perception and improved task execution capabilities in autonomous driving. LiDAR point clouds, offering accurate depth information, serve as a crucial information for this purpose. Our study focuses on key challenges in 3D target detection. To tackle the challenge of expanding the receptive field of a 3D convolutional kernel, we introduce the Dynamic Feature Fusion Module (DFFM). This module achieves adaptive expansion of the 3D convolutional kernel's receptive field, balancing the expansion with acceptable computational loads. This innovation reduces operations, expands the receptive field, and allows the model to dynamically adjust to different object requirements. Simultaneously, we identify redundant information in 3D features. Employing the Feature Selection Module (FSM) quantitatively evaluates and eliminates non-important features, achieving the separation of output box fitting and feature extraction. This innovation enables the detector to focus on critical features, resulting in model compression, reduced computational burden, and minimized candidate frame interference. Extensive experiments confirm that both DFFM and FSM not only enhance current benchmarks, particularly in small target detection, but also accelerate network performance. Importantly, these modules exhibit effective complementarity.
A Saliency Enhanced Feature Fusion based multiscale RGB-D Salient Object Detection Network
Authors: Authors: Rui Huang, Qingyi Zhao, Yan Xing, Sihua Gao, Weifeng Xu, Yuxiang Zhang, Wei Fan
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Multiscale convolutional neural network (CNN) has demonstrated remarkable capabilities in solving various vision problems. However, fusing features of different scales alwaysresults in large model sizes, impeding the application of multiscale CNNs in RGB-D saliency detection. In this paper, we propose a customized feature fusion module, called Saliency Enhanced Feature Fusion (SEFF), for RGB-D saliency detection. SEFF utilizes saliency maps of the neighboring scales to enhance the necessary features for fusing, resulting in more representative fused features. Our multiscale RGB-D saliency detector uses SEFF and processes images with three different scales. SEFF is used to fuse the features of RGB and depth images, as well as the features of decoders at different scales. Extensive experiments on five benchmark datasets have demonstrated the superiority of our method over ten SOTA saliency detectors.
Effective Intrusion Detection in Heterogeneous Internet-of-Things Networks via Ensemble Knowledge Distillation-based Federated Learning
Abstract
With the rapid development of low-cost consumer electronics and cloud computing, Internet-of-Things (IoT) devices are widely adopted for supporting next-generation distributed systems such as smart cities and industrial control systems. IoT devices are often susceptible to cyber attacks due to their open deployment environment and limited computing capabilities for stringent security controls. Hence, Intrusion Detection Systems (IDS) have emerged as one of the effective ways of securing IoT networks by monitoring and detecting abnormal activities. However, existing IDS approaches rely on centralized servers to generate behaviour profiles and detect anomalies, causing high response time and large operational costs due to communication overhead. Besides, sharing of behaviour data in an open and distributed IoT network environment may violate on-device privacy requirements. Additionally, various IoT devices tend to capture heterogeneous data, which complicates the training of behaviour models. In this paper, we introduce Federated Learning (FL) to collaboratively train a decentralized shared model of IDS, without exposing training data to others. Furthermore, we propose an effective method called Federated Learning Ensemble Knowledge Distillation (FLEKD) to mitigate the heterogeneity problems across various clients. FLEKD enables a more flexible aggregation method than conventional model fusion techniques. Experiment results on the public dataset CICIDS2019 demonstrate that the proposed approach outperforms local training and traditional FL in terms of both speed and performance and significantly improves the system's ability to detect unknown attacks. Finally, we evaluate our proposed framework's performance in three potential real-world scenarios and show FLEKD has a clear advantage in experimental results.
Claim Detection for Automated Fact-checking: A Survey on Monolingual, Multilingual and Cross-Lingual Research
Abstract
Automated fact-checking has drawn considerable attention over the past few decades due to the increase in the diffusion of misinformation on online platforms. This is often carried out as a sequence of tasks comprising (i) the detection of sentences circulating in online platforms which constitute claims needing verification, followed by (ii) the verification process of those claims. This survey focuses on the former, by discussing existing efforts towards detecting claims needing fact-checking, with a particular focus on multilingual data and methods. This is a challenging and fertile direction where existing methods are yet far from matching human performance due to the profoundly challenging nature of the issue. Especially, the dissemination of information across multiple social platforms, articulated in multiple languages and modalities demands more generalized solutions for combating misinformation. Focusing on multilingual misinformation, we present a comprehensive survey of existing multilingual claim detection research. We present state-of-the-art multilingual claim detection research categorized into three key factors of the problem, verifiability, priority, and similarity. Further, we present a detailed overview of the existing multilingual datasets along with the challenges and suggest possible future advancements.
Expert-Driven Monitoring of Operational ML Models
Authors: Authors: Joran Leest, Claudia Raibulet, Ilias Gerostathopoulos, Patricia Lago
Abstract
We propose Expert Monitoring, an approach that leverages domain expertise to enhance the detection and mitigation of concept drift in machine learning (ML) models. Our approach supports practitioners by consolidating domain expertise related to concept drift-inducing events, making this expertise accessible to on-call personnel, and enabling automatic adaptability with expert oversight.
Look, Listen and Recognise: Character-Aware Audio-Visual Subtitling
Authors: Authors: Bruno Korbar, Jaesung Huh, Andrew Zisserman
Subjects: Computer Vision and Pattern Recognition (cs.CV); Sound (cs.SD); Audio and Speech Processing (eess.AS)
Abstract
The goal of this paper is automatic character-aware subtitle generation. Given a video and a minimal amount of metadata, we propose an audio-visual method that generates a full transcript of the dialogue, with precise speech timestamps, and the character speaking identified. The key idea is to first use audio-visual cues to select a set of high-precision audio exemplars for each character, and then use these exemplars to classify all speech segments by speaker identity. Notably, the method does not require face detection or tracking. We evaluate the method over a variety of TV sitcoms, including Seinfeld, Fraiser and Scrubs. We envision this system being useful for the automatic generation of subtitles to improve the accessibility of the vast amount of videos available on modern streaming services. Project page : \url{https://www.robots.ox.ac.uk/~vgg/research/look-listen-recognise/}
NEUROSEC: FPGA-Based Neuromorphic Audio Security
Authors: Authors: Murat Isik, Hiruna Vishwamith, Yusuf Sur, Kayode Inadagbo, I. Can Dikmen
Subjects: Cryptography and Security (cs.CR); Emerging Technologies (cs.ET); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE); Sound (cs.SD); Audio and Speech Processing (eess.AS)
Abstract
Neuromorphic systems, inspired by the complexity and functionality of the human brain, have gained interest in academic and industrial attention due to their unparalleled potential across a wide range of applications. While their capabilities herald innovation, it is imperative to underscore that these computational paradigms, analogous to their traditional counterparts, are not impervious to security threats. Although the exploration of neuromorphic methodologies for image and video processing has been rigorously pursued, the realm of neuromorphic audio processing remains in its early stages. Our results highlight the robustness and precision of our FPGA-based neuromorphic system. Specifically, our system showcases a commendable balance between desired signal and background noise, efficient spike rate encoding, and unparalleled resilience against adversarial attacks such as FGSM and PGD. A standout feature of our framework is its detection rate of 94%, which, when compared to other methodologies, underscores its greater capability in identifying and mitigating threats within 5.39 dB, a commendable SNR ratio. Furthermore, neuromorphic computing and hardware security serve many sensor domains in mission-critical and privacy-preserving applications.
Out-of-Distribution Detection & Applications With Ablated Learned Temperature Energy
Authors: Authors: Will LeVine, Benjamin Pikus, Jacob Phillips, Berk Norman, Fernando Amat Gil, Sean Hendryx
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
As deep neural networks become adopted in high-stakes domains, it is crucial to be able to identify when inference inputs are Out-of-Distribution (OOD) so that users can be alerted of likely drops in performance and calibration despite high confidence. Among many others, existing methods use the following two scores to do so without training on any apriori OOD examples: a learned temperature and an energy score. In this paper we introduce Ablated Learned Temperature Energy (or "AbeT" for short), a method which combines these prior methods in novel ways with effective modifications. Due to these contributions, AbeT lowers the False Positive Rate at $95\%$ True Positive Rate (FPR@95) by $35.39\%$ in classification (averaged across all ID and OOD datasets measured) compared to state of the art without training networks in multiple stages or requiring hyperparameters or test-time backward passes. We additionally provide empirical insights as to how our model learns to distinguish between In-Distribution (ID) and OOD samples while only being explicitly trained on ID samples via exposure to misclassified ID examples at training time. Lastly, we show the efficacy of our method in identifying predicted bounding boxes and pixels corresponding to OOD objects in object detection and semantic segmentation, respectively - with an AUROC increase of $5.15\%$ in object detection and both a decrease in FPR@95 of $41.48\%$ and an increase in AUPRC of $34.20\%$ on average in semantic segmentation compared to previous state of the art.
Automated facial recognition system using deep learning for pain assessment in adults with cerebral palsy
Authors: Authors: Álvaro Sabater-Gárriz, F. Xavier Gaya-Morey, José María Buades-Rubio, Cristina Manresa Yee, Pedro Montoya, Inmaculada Riquelme
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Background: Pain assessment in individuals with neurological conditions, especially those with limited self-report ability and altered facial expressions, presents challenges. Existing measures, relying on direct observation by caregivers, lack sensitivity and specificity. In cerebral palsy, pain is a common comorbidity and a reliable evaluation protocol is crucial. Thus, having an automatic system that recognizes facial expressions could be of enormous help when diagnosing pain in this type of patient. Objectives: 1) to build a dataset of facial pain expressions in individuals with cerebral palsy, and 2) to develop an automated facial recognition system based on deep learning for pain assessment addressed to this population. Methods: Ten neural networks were trained on three pain image databases, including the UNBC-McMaster Shoulder Pain Expression Archive Database, the Multimodal Intensity Pain Dataset, and the Delaware Pain Database. Additionally, a curated dataset (CPPAIN) was created, consisting of 109 preprocessed facial pain expression images from individuals with cerebral palsy, categorized by two physiotherapists using the Facial Action Coding System observational scale. Results: InceptionV3 exhibited promising performance on the CP-PAIN dataset, achieving an accuracy of 62.67% and an F1 score of 61.12%. Explainable artificial intelligence techniques revealed consistent essential features for pain identification across models. Conclusion: This study demonstrates the potential of deep learning models for robust pain detection in populations with neurological conditions and communication disabilities. The creation of a larger dataset specific to cerebral palsy would further enhance model accuracy, offering a valuable tool for discerning subtle and idiosyncratic pain expressions. The insights gained could extend to other complex neurological conditions.
Semi-supervised segmentation of land cover images using nonlinear canonical correlation analysis with multiple features and t-SNE
Authors: Authors: Hong Wei, James Xiao, Yichao Zhang, Xia Hong
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Image segmentation is a clustering task whereby each pixel is assigned a cluster label. Remote sensing data usually consists of multiple bands of spectral images in which there exist semantically meaningful land cover subregions, co-registered with other source data such as LIDAR (LIght Detection And Ranging) data, where available. This suggests that, in order to account for spatial correlation between pixels, a feature vector associated with each pixel may be a vectorized tensor representing the multiple bands and a local patch as appropriate. Similarly, multiple types of texture features based on a pixel's local patch would also be beneficial for encoding locally statistical information and spatial variations, without necessarily labelling pixel-wise a large amount of ground truth, then training a supervised model, which is sometimes impractical. In this work, by resorting to label only a small quantity of pixels, a new semi-supervised segmentation approach is proposed. Initially, over all pixels, an image data matrix is created in high dimensional feature space. Then, t-SNE projects the high dimensional data onto 3D embedding. By using radial basis functions as input features, which use the labelled data samples as centres, to pair with the output class labels, a modified canonical correlation analysis algorithm, referred to as RBF-CCA, is introduced which learns the associated projection matrix via the small labelled data set. The associated canonical variables, obtained for the full image, are applied by k-means clustering algorithm. The proposed semi-supervised RBF-CCA algorithm has been implemented on several remotely sensed multispectral images, demonstrating excellent segmentation results.
Broiler-Net: A Deep Convolutional Framework for Broiler Behavior Analysis in Poultry Houses
Authors: Authors: Tahereh Zarrat Ehsan, Seyed Mehdi Mohtavipour
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
Detecting anomalies in poultry houses is crucial for maintaining optimal chicken health conditions, minimizing economic losses and bolstering profitability. This paper presents a novel real-time framework for analyzing chicken behavior in cage-free poultry houses to detect abnormal behaviors. Specifically, two significant abnormalities, namely inactive broiler and huddling behavior, are investigated in this study. The proposed framework comprises three key steps: (1) chicken detection utilizing a state-of-the-art deep learning model, (2) tracking individual chickens across consecutive frames with a fast tracker module, and (3) detecting abnormal behaviors within the video stream. Experimental studies are conducted to evaluate the efficacy of the proposed algorithm in accurately assessing chicken behavior. The results illustrate that our framework provides a precise and efficient solution for real-time anomaly detection, facilitating timely interventions to maintain chicken health and enhance overall productivity on poultry farms. Github: https://github.com/TaherehZarratEhsan/Chicken-Behavior-Analysis
Is Your Kettle Smarter Than a Hacker? A Scalable Tool for Assessing Replay Attack Vulnerabilities on Consumer IoT Devices
Authors: Authors: Sara Lazzaro, Vincenzo De Angelis, Anna Maria Mandalari, Francesco Buccafurri
Abstract
Consumer Internet of Things (IoT) devices often leverage the local network to communicate with the corresponding companion app or other devices. This has benefits in terms of efficiency since it offloads the cloud. ENISA and NIST security guidelines underscore the importance of enabling default local communication for safety and reliability. Indeed, an IoT device should continue to function in case the cloud connection is not available. While the security of cloud-device connections is typically strengthened through the usage of standard protocols, local connectivity security is frequently overlooked. Neglecting the security of local communication opens doors to various threats, including replay attacks. In this paper, we investigate this class of attacks by designing a systematic methodology for automatically testing IoT devices vulnerability to replay attacks. Specifically, we propose a tool, named REPLIOT, able to test whether a replay attack is successful or not, without prior knowledge of the target devices. We perform thousands of automated experiments using popular commercial devices spanning various vendors and categories. Notably, our study reveals that among these devices, 51% of them do not support local connectivity, thus they are not compliant with the reliability and safety requirements of the ENISA/NIST guidelines. We find that 75% of the remaining devices are vulnerable to replay attacks with REPLIOT having a detection accuracy of 0.98-1. Finally, we investigate the possible causes of this vulnerability, discussing possible mitigation strategies.
Programmable EM Sensor Array for Golden-Model Free Run-time Trojan Detection and Localization
Authors: Authors: Hanqiu Wang, Max Panoff, Zihao Zhan, Shuo Wang, Christophe Bobda, Domenic Forte
Subjects: Cryptography and Security (cs.CR); Signal Processing (eess.SP)
Abstract
Side-channel analysis has been proven effective at detecting hardware Trojans in integrated circuits (ICs). However, most detection techniques rely on large external probes and antennas for data collection and require a long measurement time to detect Trojans. Such limitations make these techniques impractical for run-time deployment and ineffective in detecting small Trojans with subtle side-channel signatures. To overcome these challenges, we propose a Programmable Sensor Array (PSA) for run-time hardware Trojan detection, localization, and identification. PSA is a tampering-resilient integrated on-chip magnetic field sensor array that can be re-programmed to change the sensors' shape, size, and location. Using PSA, EM side-channel measurement results collected from sensors at different locations on an IC can be analyzed to localize and identify the Trojan. The PSA has better performance than conventional external magnetic probes and state-of-the-art on-chip single-coil magnetic field sensors. We fabricated an AES-128 test chip with four AES Hardware Trojans. They were successfully detected, located, and identified with the proposed on-chip PSA within 10 milliseconds using our proposed cross-domain analysis.
OK-Robot: What Really Matters in Integrating Open-Knowledge Models for Robotics
Authors: Authors: Peiqi Liu, Yaswanth Orru, Chris Paxton, Nur Muhammad Mahi Shafiullah, Lerrel Pinto
Abstract
Remarkable progress has been made in recent years in the fields of vision, language, and robotics. We now have vision models capable of recognizing objects based on language queries, navigation systems that can effectively control mobile systems, and grasping models that can handle a wide range of objects. Despite these advancements, general-purpose applications of robotics still lag behind, even though they rely on these fundamental capabilities of recognition, navigation, and grasping. In this paper, we adopt a systems-first approach to develop a new Open Knowledge-based robotics framework called OK-Robot. By combining Vision-Language Models (VLMs) for object detection, navigation primitives for movement, and grasping primitives for object manipulation, OK-Robot offers a integrated solution for pick-and-drop operations without requiring any training. To evaluate its performance, we run OK-Robot in 10 real-world home environments. The results demonstrate that OK-Robot achieves a 58.5% success rate in open-ended pick-and-drop tasks, representing a new state-of-the-art in Open Vocabulary Mobile Manipulation (OVMM) with nearly 1.8x the performance of prior work. On cleaner, uncluttered environments, OK-Robot's performance increases to 82%. However, the most important insight gained from OK-Robot is the critical role of nuanced details when combining Open Knowledge systems like VLMs with robotic modules. Videos of our experiments are available on our website: https://ok-robot.github.io
Keyword: face recognition
TetraLoss: Improving the Robustness of Face Recognition against Morphing Attacks
Authors: Authors: Mathias Ibsen, Lázaro J. González-Soler, Christian Rathgeb, Christoph Busch
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Face recognition systems are widely deployed in high-security applications such as for biometric verification at border controls. Despite their high accuracy on pristine data, it is well-known that digital manipulations, such as face morphing, pose a security threat to face recognition systems. Malicious actors can exploit the facilities offered by the identity document issuance process to obtain identity documents containing morphed images. Thus, subjects who contributed to the creation of the morphed image can with high probability use the identity document to bypass automated face recognition systems. In recent years, no-reference (i.e., single image) and differential morphing attack detectors have been proposed to tackle this risk. These systems are typically evaluated in isolation from the face recognition system that they have to operate jointly with and do not consider the face recognition process. Contrary to most existing works, we present a novel method for adapting deep learning-based face recognition systems to be more robust against face morphing attacks. To this end, we introduce TetraLoss, a novel loss function that learns to separate morphed face images from its contributing subjects in the embedding space while still preserving high biometric verification performance. In a comprehensive evaluation, we show that the proposed method can significantly enhance the original system while also significantly outperforming other tested baseline methods.
Keyword: augmentation
Does Using ChatGPT Result in Human Cognitive Augmentation?
Abstract
Human cognitive performance is enhanced by the use of tools. For example, a human can produce a much greater, and more accurate, volume of mathematical calculation in a unit of time using a calculator or a spreadsheet application on a computer. Such tools have taken over the burden of lower level cognitive grunt work but the human still serves the role of the expert performing higher level thinking and reasoning. Recently, however, unsupervised, deep, machine learning has produced cognitive systems able to outperform humans in several domains. When humans use these tools in a human cog ensemble, the cognitive ability of the human is augmented. In some cases, even non experts can achieve, and even exceed, the performance of experts in a particular domain, synthetic expertise. A new cognitive system, ChatGPT, has burst onto the scene during the past year. This paper investigates human cognitive augmentation due to using ChatGPT by presenting the results of two experiments comparing responses created using ChatGPT with results created not using ChatGPT. We find using ChatGPT does not always result in cognitive augmentation and does not yet replace human judgement, discernment, and evaluation in certain types of tasks. In fact, ChatGPT was observed to result in misleading users resulting in negative cognitive augmentation.
End-to-End Argument Mining over Varying Rhetorical Structures
Abstract
Rhetorical Structure Theory implies no single discourse interpretation of a text, and the limitations of RST parsers further exacerbate inconsistent parsing of similar structures. Therefore, it is important to take into account that the same argumentative structure can be found in semantically similar texts with varying rhetorical structures. In this work, the differences between paraphrases within the same argument scheme are evaluated from a rhetorical perspective. The study proposes a deep dependency parsing model to assess the connection between rhetorical and argument structures. The model utilizes rhetorical relations; RST structures of paraphrases serve as training data augmentations. The method allows for end-to-end argumentation analysis using a rhetorical tree instead of a word sequence. It is evaluated on the bilingual Microtexts corpus, and the first results on fully-fledged argument parsing for the Russian version of the corpus are reported. The results suggest that argument mining can benefit from multiple variants of discourse structure.
Closing the Gap between TD Learning and Supervised Learning -- A Generalisation Point of View
Authors: Authors: Raj Ghugare, Matthieu Geist, Glen Berseth, Benjamin Eysenbach
Abstract
Some reinforcement learning (RL) algorithms can stitch pieces of experience to solve a task never seen before during training. This oft-sought property is one of the few ways in which RL methods based on dynamic-programming differ from RL methods based on supervised-learning (SL). Yet, certain RL methods based on off-the-shelf SL algorithms achieve excellent results without an explicit mechanism for stitching; it remains unclear whether those methods forgo this important stitching property. This paper studies this question for the problems of achieving a target goal state and achieving a target return value. Our main result is to show that the stitching property corresponds to a form of combinatorial generalization: after training on a distribution of (state, goal) pairs, one would like to evaluate on (state, goal) pairs not seen together in the training data. Our analysis shows that this sort of generalization is different from i.i.d. generalization. This connection between stitching and generalisation reveals why we should not expect SL-based RL methods to perform stitching, even in the limit of large datasets and models. Based on this analysis, we construct new datasets to explicitly test for this property, revealing that SL-based methods lack this stitching property and hence fail to perform combinatorial generalization. Nonetheless, the connection between stitching and combinatorial generalisation also suggests a simple remedy for improving generalisation in SL: data augmentation. We propose a temporal data augmentation and demonstrate that adding it to SL-based methods enables them to successfully complete tasks not seen together during training. On a high level, this connection illustrates the importance of combinatorial generalization for data efficiency in time-series data beyond tasks beyond RL, like audio, video, or text.
Adversarial Augmentation Training Makes Action Recognition Models More Robust to Realistic Video Distribution Shifts
Authors: Authors: Kiyoon Kim, Shreyank N Gowda, Panagiotis Eustratiadis, Antreas Antoniou, Robert B Fisher
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Despite recent advances in video action recognition achieving strong performance on existing benchmarks, these models often lack robustness when faced with natural distribution shifts between training and test data. We propose two novel evaluation methods to assess model resilience to such distribution disparity. One method uses two different datasets collected from different sources and uses one for training and validation, and the other for testing. More precisely, we created dataset splits of HMDB-51 or UCF-101 for training, and Kinetics-400 for testing, using the subset of the classes that are overlapping in both train and test datasets. The other proposed method extracts the feature mean of each class from the target evaluation dataset's training data (i.e. class prototype) and estimates test video prediction as a cosine similarity score between each sample to the class prototypes of each target class. This procedure does not alter model weights using the target dataset and it does not require aligning overlapping classes of two different datasets, thus is a very efficient method to test the model robustness to distribution shifts without prior knowledge of the target distribution. We address the robustness problem by adversarial augmentation training - generating augmented views of videos that are "hard" for the classification model by applying gradient ascent on the augmentation parameters - as well as "curriculum" scheduling the strength of the video augmentations. We experimentally demonstrate the superior performance of the proposed adversarial augmentation approach over baselines across three state-of-the-art action recognition models - TSM, Video Swin Transformer, and Uniformer. The presented work provides critical insight into model robustness to distribution shifts and presents effective techniques to enhance video action recognition performance in a real-world deployment.
Towards Better Inclusivity: A Diverse Tweet Corpus of English Varieties
Authors: Authors: Nhi Pham, Lachlan Pham, Adam L. Meyers
Subjects: Computation and Language (cs.CL); Computers and Society (cs.CY)
Abstract
The prevalence of social media presents a growing opportunity to collect and analyse examples of English varieties. Whilst usage of these varieties was - and, in many cases, still is - used only in spoken contexts or hard-to-access private messages, social media sites like Twitter provide a platform for users to communicate informally in a scrapeable format. Notably, Indian English (Hinglish), Singaporean English (Singlish), and African-American English (AAE) can be commonly found online. These varieties pose a challenge to existing natural language processing (NLP) tools as they often differ orthographically and syntactically from standard English for which the majority of these tools are built. NLP models trained on standard English texts produced biased outcomes for users of underrepresented varieties. Some research has aimed to overcome the inherent biases caused by unrepresentative data through techniques like data augmentation or adjusting training models. We aim to address the issue of bias at its root - the data itself. We curate a dataset of tweets from countries with high proportions of underserved English variety speakers, and propose an annotation framework of six categorical classifications along a pseudo-spectrum that measures the degree of standard English and that thereby indirectly aims to surface the manifestations of English varieties in these tweets. Following best annotation practices, our growing corpus features 170,800 tweets taken from 7 countries, labeled by annotators who are from those countries and can communicate in regionally-dominant varieties of English. Our corpus highlights the accuracy discrepancies in pre-trained language identifiers between western English and non-western (i.e., less standard) English varieties. We hope to contribute to the growing literature identifying and reducing the implicit demographic discrepancies in NLP.
Blinded by Generated Contexts: How Language Models Merge Generated and Retrieved Contexts for Open-Domain QA?
Abstract
While auxiliary information has become a key to enhance Large Language Models (LLMs), relatively little is known about how well LLMs merge these contexts, specifically generated and retrieved. To study this, we formulate a task specifically designed to identify whether the answers, derived from the integration of generated and retrieved contexts, are attributed to either generated or retrieved contexts. To support this task, we develop a methodology to construct datasets with conflicting contexts, where each question is paired with both generated and retrieved contexts, yet only one of them contains the correct answer. Our experiments reveal a significant bias in LLMs towards generated contexts, as evidenced across state-of-the-art open (Llama2-7b/13b) and closed (GPT 3.5/4) systems. We further identify two key factors contributing to this bias: i) Contexts generated by LLMs typically show greater similarity to the questions, increasing their likelihood of selection; ii) The segmentation process used in retrieved contexts disrupts their completeness, thereby hindering their full utilization in LLMs. Our analysis enhances the understanding of how LLMs merge diverse contexts, offering valuable insights for advancing current augmentation methods for LLMs.
Keyword: detection
One Step Learning, One Step Review
Analysis and Detection of Multilingual Hate Speech Using Transformer Based Deep Learning
Exploring Highly Quantised Neural Networks for Intrusion Detection in Automotive CAN
Low Complexity Turbo SIC-MMSE Detection for Orthogonal Time Frequency Space Modulation
The Best Ends for the Best Means: Ethical Concerns in App Reviews
LLM4Fuzz: Guided Fuzzing of Smart Contracts with Large Language Models
Stability Plasticity Decoupled Fine-tuning For Few-shot end-to-end Object Detection
BinaryAI: Binary Software Composition Analysis via Intelligent Binary Source Code Matching
Pixel-Wise Recognition for Holistic Surgical Scene Understanding
Unfair TOS: An Automated Approach using Customized BERT
TreeMIL: A Multi-instance Learning Framework for Time Series Anomaly Detection with Inexact Supervision
Word-Level ASR Quality Estimation for Efficient Corpus Sampling and Post-Editing through Analyzing Attentions of a Reference-Free Metric
DACR: Distribution-Augmented Contrastive Reconstruction for Time-Series Anomaly Detection
Source Detection in Networks using the Stationary Distribution of a Markov Chain
ANNA: A Deep Learning Based Dataset in Heterogeneous Traffic for Autonomous Vehicles
A Fast Parallel Approach for Neighborhood-based Link Prediction by Disregarding Large Hubs
MapChange: Enhancing Semantic Change Detection with Temporal-Invariant Historical Maps Based on Deep Triplet Network
Edge-Enabled Real-time Railway Track Segmentation
Thermal Image Calibration and Correction using Unpaired Cycle-Consistent Adversarial Networks
A Survey on African Computer Vision Datasets, Topics and Researchers
EK-Net:Real-time Scene Text Detection with Expand Kernel Distance
MsSVT++: Mixed-scale Sparse Voxel Transformer with Center Voting for 3D Object Detection
Unraveling Attacks in Machine Learning-based IoT Ecosystems: A Survey and the Open Libraries Behind Them
Detecting Out-of-Distribution Samples via Conditional Distribution Entropy with Optimal Transport
GI-PIP: Do We Require Impractical Auxiliary Dataset for Gradient Inversion Attacks?
Identity-Driven Multimedia Forgery Detection via Reference Assistance
Concealed Object Segmentation with Hierarchical Coherence Modeling
Deep Learning for Computer Vision based Activity Recognition and Fall Detection of the Elderly: a Systematic Review
Symbrain: A large-scale dataset of MRI images for neonatal brain symmetry analysis
MInD: Improving Multimodal Sentiment Analysis via Multimodal Information Disentanglement
Rethinking Centered Kernel Alignment in Knowledge Distillation
End-to-end Multi-Instance Robotic Reaching from Monocular Vision
A Review of Physics-Informed Machine Learning Methods with Applications to Condition Monitoring and Anomaly Detection
Detect-Order-Construct: A Tree Construction based Approach for Hierarchical Document Structure Analysis
Large receptive field strategy and important feature extraction strategy in 3D object detection
A Saliency Enhanced Feature Fusion based multiscale RGB-D Salient Object Detection Network
Effective Intrusion Detection in Heterogeneous Internet-of-Things Networks via Ensemble Knowledge Distillation-based Federated Learning
Claim Detection for Automated Fact-checking: A Survey on Monolingual, Multilingual and Cross-Lingual Research
Expert-Driven Monitoring of Operational ML Models
Look, Listen and Recognise: Character-Aware Audio-Visual Subtitling
NEUROSEC: FPGA-Based Neuromorphic Audio Security
Out-of-Distribution Detection & Applications With Ablated Learned Temperature Energy
Automated facial recognition system using deep learning for pain assessment in adults with cerebral palsy
Semi-supervised segmentation of land cover images using nonlinear canonical correlation analysis with multiple features and t-SNE
Broiler-Net: A Deep Convolutional Framework for Broiler Behavior Analysis in Poultry Houses
Is Your Kettle Smarter Than a Hacker? A Scalable Tool for Assessing Replay Attack Vulnerabilities on Consumer IoT Devices
Programmable EM Sensor Array for Golden-Model Free Run-time Trojan Detection and Localization
OK-Robot: What Really Matters in Integrating Open-Knowledge Models for Robotics
Keyword: face recognition
TetraLoss: Improving the Robustness of Face Recognition against Morphing Attacks
Keyword: augmentation
Does Using ChatGPT Result in Human Cognitive Augmentation?
End-to-End Argument Mining over Varying Rhetorical Structures
Closing the Gap between TD Learning and Supervised Learning -- A Generalisation Point of View
Adversarial Augmentation Training Makes Action Recognition Models More Robust to Realistic Video Distribution Shifts
Towards Better Inclusivity: A Diverse Tweet Corpus of English Varieties
Blinded by Generated Contexts: How Language Models Merge Generated and Retrieved Contexts for Open-Domain QA?