Abstract
We propose an approach for curating multimodal data that we used for our entry in the 2023 DataComp competition filtering track. Our technique combines object detection and weak supervision-based ensembling. In the first of two steps in our approach, we employ an out-of-the-box zero-shot object detection model to extract granular information and produce a variety of filter designs. In the second step, we employ weak supervision to ensemble filtering rules. This approach results in a 4% performance improvement when compared to the best-performing baseline, producing the top-ranking position in the small scale track at the time of writing. Furthermore, in the medium scale track, we achieve a noteworthy 4.2% improvement over the baseline by simply ensembling existing baselines with weak supervision.
A Lightweight FPGA-based IDS-ECU Architecture for Automotive CAN
Abstract
Recent years have seen an exponential rise in complex software-driven functionality in vehicles, leading to a rising number of electronic control units (ECUs), network capabilities, and interfaces. These expanded capabilities also bring-in new planes of vulnerabilities making intrusion detection and management a critical capability; however, this can often result in more ECUs and network elements due to the high computational overheads. In this paper, we present a consolidated ECU architecture incorporating an Intrusion Detection System (IDS) for Automotive Controller Area Network (CAN) along with traditional ECU functionality on an off-the-shelf hybrid FPGA device, with near-zero overhead for the ECU functionality. We propose two quantised multi-layer perceptrons (QMLP's) as isolated IDSs for detecting a range of attack vectors including Denial-of-Service, Fuzzing and Spoofing, which are accelerated using off-the-shelf deep-learning processing unit (DPU) IP block from Xilinx, operating fully transparently to the software on the ECU. The proposed models achieve the state-of-the-art classification accuracy for all the attacks, while we observed a 15x reduction in power consumption when compared against the GPU-based implementation of the same models quantised using Nvidia libraries. We also achieved a 2.3x speed up in per-message processing latency (at 0.24 ms from the arrival of a CAN message) to meet the strict end-to-end latency on critical CAN nodes and a 2.6x reduction in power consumption for inference when compared to the state-of-the-art IDS models on embedded IDS and loosely coupled IDS accelerators (GPUs) discussed in the literature.
Quantised Neural Network Accelerators for Low-Power IDS in Automotive Networks
Abstract
In this paper, we explore low-power custom quantised Multi-Layer Perceptrons (MLPs) as an Intrusion Detection System (IDS) for automotive controller area network (CAN). We utilise the FINN framework from AMD/Xilinx to quantise, train and generate hardware IP of our MLP to detect denial of service (DoS) and fuzzying attacks on CAN network, using ZCU104 (XCZU7EV) FPGA as our target ECU architecture with integrated IDS capabilities. Our approach achieves significant improvements in latency (0.12 ms per-message processing latency) and inference energy consumption (0.25 mJ per inference) while achieving similar classification performance as state-of-the-art approaches in the literature.
Machine learning-based network intrusion detection for big and imbalanced data using oversampling, stacking feature embedding and feature extraction
Authors: Authors: Md. Alamin Talukder, Md. Manowarul Islam, Md Ashraf Uddin, Khondokar Fida Hasan, Selina Sharmin, Salem A. Alyami, Mohammad Ali Moni
Subjects: Cryptography and Security (cs.CR); Machine Learning (cs.LG)
Abstract
Cybersecurity has emerged as a critical global concern. Intrusion Detection Systems (IDS) play a critical role in protecting interconnected networks by detecting malicious actors and activities. Machine Learning (ML)-based behavior analysis within the IDS has considerable potential for detecting dynamic cyber threats, identifying abnormalities, and identifying malicious conduct within the network. However, as the number of data grows, dimension reduction becomes an increasingly difficult task when training ML models. Addressing this, our paper introduces a novel ML-based network intrusion detection model that uses Random Oversampling (RO) to address data imbalance and Stacking Feature Embedding based on clustering results, as well as Principal Component Analysis (PCA) for dimension reduction and is specifically designed for large and imbalanced datasets. This model's performance is carefully evaluated using three cutting-edge benchmark datasets: UNSW-NB15, CIC-IDS-2017, and CIC-IDS-2018. On the UNSW-NB15 dataset, our trials show that the RF and ET models achieve accuracy rates of 99.59% and 99.95%, respectively. Furthermore, using the CIC-IDS2017 dataset, DT, RF, and ET models reach 99.99% accuracy, while DT and RF models obtain 99.94% accuracy on CIC-IDS2018. These performance results continuously outperform the state-of-art, indicating significant progress in the field of network intrusion detection. This achievement demonstrates the efficacy of the suggested methodology, which can be used practically to accurately monitor and identify network traffic intrusions, thereby blocking possible threats.
Fine-tuning Large Language Models for Multigenerator, Multidomain, and Multilingual Machine-Generated Text Detection
Abstract
SemEval-2024 Task 8 introduces the challenge of identifying machine-generated texts from diverse Large Language Models (LLMs) in various languages and domains. The task comprises three subtasks: binary classification in monolingual and multilingual (Subtask A), multi-class classification (Subtask B), and mixed text detection (Subtask C). This paper focuses on Subtask A & B. Each subtask is supported by three datasets for training, development, and testing. To tackle this task, two methods: 1) using traditional machine learning (ML) with natural language preprocessing (NLP) for feature extraction, and 2) fine-tuning LLMs for text classification. The results show that transformer models, particularly LoRA-RoBERTa, exceed traditional ML methods in effectiveness, with majority voting being particularly effective in multilingual contexts for identifying machine-generated texts.
ACS: Concurrent Kernel Execution on Irregular, Input-Dependent Computational Graphs
Authors: Authors: Sankeerth Durvasula, Adrian Zhao, Raymond Kiguru, Yushi Guan, Zhonghan Chen, Nandita Vijaykumar
Abstract
GPUs are widely used to accelerate many important classes of workloads today. However, we observe that several important emerging classes of workloads, including simulation engines for deep reinforcement learning and dynamic neural networks, are unable to fully utilize the massive parallelism that GPUs offer. These applications tend to have kernels that are small in size, i.e., have few thread blocks that do not saturate compute resources. Executing independent kernels concurrently is a promising approach to improve parallelism and utilization. However, this inter-kernel concurrency is difficult to leverage in such workloads with existing approaches: First, the inter-kernel dependencies and computational graph are input-dependent and vary each time the application is executed. Second, the computational graphs tend to be irregular, requiring fine-grain scheduling and synchronization; thus incurring significant synchronization overheads if kernel execution is parallelized. In this work, we propose ACS, a framework that enables lightweight detection of inter-kernel dependencies and low overhead kernel scheduling at runtime. The key idea behind ACS is to perform inter-kernel dependency checks for a small window of kernels at runtime, similar to out-of order instruction scheduling. This enables concurrent execution of kernels in applications whose computational graphs are input dependent and require fine-grained scheduling. We propose ACS-SW, a software-only open-source implementation of ACS and ACS-HW, a hardware-software cooperative implementation. ACS-HW further reduces synchronization overheads by reducing communication between the CPU and GPU. We evaluate ACS for deep RL simulation and dynamic DNNs on both real hardware and a GPU simulator. We demonstrate speedups of up to 2.19x (1.56x on average) by improving GPU utilization with concurrent kernel execution.
Learning Recovery Strategies for Dynamic Self-healing in Reactive Systems
Authors: Authors: Mateo Sanabria, Ivana Dusparic, Nicolas Cardozo
Subjects: Distributed, Parallel, and Cluster Computing (cs.DC); Software Engineering (cs.SE)
Abstract
Self-healing systems depend on following a set of predefined instructions to recover from a known failure state. Failure states are generally detected based on domain specific specialized metrics. Failure fixes are applied at predefined application hooks that are not sufficiently expressive to manage different failure types. Self-healing is usually applied in the context of distributed systems, where the detection of failures is constrained to communication problems, and resolution strategies often consist of replacing complete components. Our proposal targets complex reactive systems, defining monitors as predicates specifying satisfiability conditions of system properties. Such monitors are functionally expressive and can be defined at run time to detect failure states at any execution point. Once failure states are detected, we use a Reinforcement Learning-based technique to learn a recovery strategy based on users' corrective sequences. Finally, to execute the learned strategies, we extract them as COP variations that activate dynamically whenever the failure state is detected, overwriting the base system behavior with the recovery strategy for that state. We validate the feasibility and effectiveness of our framework through a prototypical reactive application for tracking mouse movements, and the DeltaIoT exemplar for self-healing systems. Our results demonstrate that with just the definition of monitors, the system is effective in detecting and recovering from failures between 55%-92% of the cases in the first application, and at par with the predefined strategies in the second application.
Patch2QL: Discover Cognate Defects in Open Source Software Supply Chain With Auto-generated Static Analysis Rules
Authors: Authors: Fuwei Wang
Subjects: Cryptography and Security (cs.CR); Software Engineering (cs.SE)
Abstract
In the open source software (OSS) ecosystem, there exists a complex software supply chain, where developers upstream and downstream widely borrow and reuse code. This results in the widespread occurrence of recurring defects, missing fixes, and propagation issues. These are collectively referred to as cognate defects, and their scale and threats have not received extensive attention and systematic research. Software composition analysis and code clone detection methods are unable to cover the various variant issues in the supply chain scenario, while code static analysis, or static application security testing (SAST) techniques struggle to target specific defects. In this paper, we propose a novel technique for detecting cognate defects in OSS through the automatic generation of SAST rules. Specifically, it extracts key syntax and semantic information from pre- and post-patch versions of code through structural comparison and control flow to data flow analysis, and generates rules that matches these key elements. We have implemented a prototype tool called Patch2QL and applied it to fundamental OSS in C/C++. In experiments, we discovered 7 new vulnerabilities with medium to critical severity in the most popular upstream software, as well as numerous potential security issues. When analyzing downstream projects in the supply chain, we found a significant number of representative cognate defects, clarifying the threat posed by this issue. Additionally, compared to general-purpose SAST and signature-based mechanisms, the generated rules perform better at discover all variants of cognate defects.
NIV-SSD: Neighbor IoU-Voting Single-Stage Object Detector From Point Cloud
Authors: Authors: Shuai Liu, Di Wang, Quan Wang, Kai Huang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Previous single-stage detectors typically suffer the misalignment between localization accuracy and classification confidence. To solve the misalignment problem, we introduce a novel rectification method named neighbor IoU-voting (NIV) strategy. Typically, classification and regression are treated as separate branches, making it challenging to establish a connection between them. Consequently, the classification confidence cannot accurately reflect the regression quality. NIV strategy can serve as a bridge between classification and regression branches by calculating two types of statistical data from the regression output to correct the classification confidence. Furthermore, to alleviate the imbalance of detection accuracy for complete objects with dense points (easy objects) and incomplete objects with sparse points (difficult objects), we propose a new data augmentation scheme named object resampling. It undersamples easy objects and oversamples difficult objects by randomly transforming part of easy objects into difficult objects. Finally, combining the NIV strategy and object resampling augmentation, we design an efficient single-stage detector termed NIV-SSD. Extensive experiments on several datasets indicate the effectiveness of the NIV strategy and the competitive performance of the NIV-SSD detector. The code will be available at https://github.com/Say2L/NIV-SSD.
Self-supervised Learning of LiDAR 3D Point Clouds via 2D-3D Neural Calibration
Abstract
This paper introduces a novel self-supervised learning framework for enhancing 3D perception in autonomous driving scenes. Specifically, our approach, named NCLR, focuses on 2D-3D neural calibration, a novel pretext task that estimates the rigid transformation aligning camera and LiDAR coordinate systems. First, we propose the learnable transformation alignment to bridge the domain gap between image and point cloud data, converting features into a unified representation space for effective comparison and matching. Second, we identify the overlapping area between the image and point cloud with the fused features. Third, we establish dense 2D-3D correspondences to estimate the rigid transformation. The framework not only learns fine-grained matching from points to pixels but also achieves alignment of the image and point cloud at a holistic level, understanding their relative pose. We demonstrate NCLR's efficacy by applying the pre-trained backbone to downstream tasks, such as LiDAR-based 3D semantic segmentation, object detection, and panoptic segmentation. Comprehensive experiments on various datasets illustrate the superiority of NCLR over existing self-supervised methods. The results confirm that joint learning from different modalities significantly enhances the network's understanding abilities and effectiveness of learned representation. Code will be available at \url{https://github.com/Eaphan/NCLR}.
Comparing Human-Centered Language Modeling: Is it Better to Model Groups, Individual Traits, or Both?
Authors: Authors: Nikita Soni, Niranjan Balasubramanian, H. Andrew Schwartz, Dirk Hovy
Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Abstract
Natural language processing has made progress in incorporating human context into its models, but whether it is more effective to use group-wise attributes (e.g., over-45-year-olds) or model individuals remains open. Group attributes are technically easier but coarse: not all 45-year-olds write the same way. In contrast, modeling individuals captures the complexity of each person's identity. It allows for a more personalized representation, but we may have to model an infinite number of users and require data that may be impossible to get. We compare modeling human context via group attributes, individual users, and combined approaches. Combining group and individual features significantly benefits user-level regression tasks like age estimation or personality assessment from a user's documents. Modeling individual users significantly improves the performance of single document-level classification tasks like stance and topic detection. We also find that individual-user modeling does well even without user's historical data.
On the Fundamental Tradeoff of Joint Communication and Quickest Change Detection
Authors: Authors: Daewon Seo, Sung Hoon Lim
Subjects: Information Theory (cs.IT); Signal Processing (eess.SP)
Abstract
In this work, we take the initiative in studying the fundamental tradeoff between communication and quickest change detection (QCD) under an integrated sensing and communication setting. We formally establish a joint communication and sensing problem for quickest change detection. Then, by utilizing constant subblock-composition codes and a modified QuSum detection rule, which we call subblock QuSum (SQS), we provide an inner bound on the fundamental tradeoff between communication rate and change point detection delay in the asymptotic regime of vanishing false alarm rate. We further provide a partial converse that matches our inner bound for a certain class of codes. This implies that the SQS detection strategy is asymptotically optimal for our codes as the false alarm rate constraint vanishes. We also present some canonical examples of the tradeoff region for a binary channel, a scalar Gaussian channel, and a MIMO Gaussian channel.
Small Language Model Meets with Reinforced Vision Vocabulary
Abstract
Playing Large Vision Language Models (LVLMs) in 2023 is trendy among the AI community. However, the relatively large number of parameters (more than 7B) of popular LVLMs makes it difficult to train and deploy on consumer GPUs, discouraging many researchers with limited resources. Imagine how cool it would be to experience all the features of current LVLMs on an old GTX1080ti (our only game card). Accordingly, we present Vary-toy in this report, a small-size Vary along with Qwen-1.8B as the base ``large'' language model. In Vary-toy, we introduce an improved vision vocabulary, allowing the model to not only possess all features of Vary but also gather more generality. Specifically, we replace negative samples of natural images with positive sample data driven by object detection in the procedure of generating vision vocabulary, more sufficiently utilizing the capacity of the vocabulary network and enabling it to efficiently encode visual information corresponding to natural objects. For experiments, Vary-toy can achieve 65.6% ANLS on DocVQA, 59.1% accuracy on ChartQA, 88.1% accuracy on RefCOCO, and 29% on MMVet. The code will be publicly available on the homepage.
Abstract
Facial expression recognition (FER) models are typically trained on datasets with a fixed number of seven basic classes. However, recent research works point out that there are far more expressions than the basic ones. Thus, when these models are deployed in the real world, they may encounter unknown classes, such as compound expressions that cannot be classified into existing basic classes. To address this issue, we propose the open-set FER task for the first time. Though there are many existing open-set recognition methods, we argue that they do not work well for open-set FER because FER data are all human faces with very small inter-class distances, which makes the open-set samples very similar to close-set samples. In this paper, we are the first to transform the disadvantage of small inter-class distance into an advantage by proposing a new way for open-set FER. Specifically, we find that small inter-class distance allows for sparsely distributed pseudo labels of open-set samples, which can be viewed as symmetric noisy labels. Based on this novel observation, we convert the open-set FER to a noisy label detection problem. We further propose a novel method that incorporates attention map consistency and cycle training to detect the open-set samples. Extensive experiments on various FER datasets demonstrate that our method clearly outperforms state-of-the-art open-set recognition methods by large margins. Code is available at https://github.com/zyh-uaiaaaa.
Detecting and recognizing characters in Greek papyri with YOLOv8, DeiT and SimCLR
Authors: Authors: Robert Turnbull, Evelyn Mannix
Subjects: Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI)
Abstract
The capacity to isolate and recognize individual characters from facsimile images of papyrus manuscripts yields rich opportunities for digital analysis. For this reason the `ICDAR 2023 Competition on Detection and Recognition of Greek Letters on Papyri' was held as part of the 17th International Conference on Document Analysis and Recognition. This paper discusses our submission to the competition. We used an ensemble of YOLOv8 models to detect and classify individual characters and employed two different approaches for refining the character predictions, including a transformer based DeiT approach and a ResNet-50 model trained on a large corpus of unlabelled data using SimCLR, a self-supervised learning method. Our submission won the recognition challenge with a mAP of 42.2%, and was runner-up in the detection challenge with a mean average precision (mAP) of 51.4%. At the more relaxed intersection over union threshold of 0.5, we achieved the highest mean average precision and mean average recall results for both detection and classification. We ran our prediction pipeline on more than 4,500 images from the Oxyrhynchus Papyri to illustrate the utility of our approach, and we release the results publicly in multiple formats.
Multi-Party Private Set Intersection: A Circuit-Based Protocol with Jaccard Similarity for Secure and Efficient Anomaly Detection in Network Traffic
Authors: Authors: Jiuheng Su, Zhili Chen, Xiaomin Yang
Abstract
We present a new circuit-based protocol for multi-party private set intersection (PSI) that allows m parties to compute the intersection of their datasets without revealing any additional information about the items outside the intersection. Building upon the two-party Sort-Compare-Shuffle (SCS) protocol, we seamlessly extend it to a multi-party setting. Demonstrating its practicality through implementation, our protocol exhibits acceptable performance. Specifically, with 7 parties, each possessing a set size of 2^{12}, our protocol completes in just 19 seconds. Moreover, circuit-based protocols like ours have an advantage over using custom protocols to perform more complex computation. We substantiate this advantage by incorporating a module for calculating the Jaccard similarity metric of the private sets which can be used in the application domain of network traffic analysis for anomaly detection. This extension showcases the versatility of our protocol beyond set intersection computations, demonstrating its efficacy in preserving privacy while efficiently identifying abnormal patterns in network flow.
Blind Channel Estimation and Joint Symbol Detection with Data-Driven Factor Graphs
Abstract
We investigate the application of the factor graph framework for blind joint channel estimation and symbol detection on time-variant linear inter-symbol interference channels. In particular, we consider the expectation maximization (EM) algorithm for maximum likelihood estimation, which typically suffers from high complexity as it requires the computation of the symbol-wise posterior distributions in every iteration. We address this issue by efficiently approximating the posteriors using the belief propagation (BP) algorithm on a suitable factor graph. By interweaving the iterations of BP and EM, the detection complexity can be further reduced to a single BP iteration per EM step. In addition, we propose a data-driven version of our algorithm that introduces momentum in the BP updates and learns a suitable EM parameter update schedule, thereby significantly improving the performance-complexity tradeoff with a few offline training samples. Our numerical experiments demonstrate the excellent performance of the proposed blind detector and show that it even outperforms coherent BP detection in high signal-to-noise scenarios.
Pragmatic Communication in Multi-Agent Collaborative Perception
Abstract
Collaborative perception allows each agent to enhance its perceptual abilities by exchanging messages with others. It inherently results in a trade-off between perception ability and communication costs. Previous works transmit complete full-frame high-dimensional feature maps among agents, resulting in substantial communication costs. To promote communication efficiency, we propose only transmitting the information needed for the collaborator's downstream task. This pragmatic communication strategy focuses on three key aspects: i) pragmatic message selection, which selects task-critical parts from the complete data, resulting in spatially and temporally sparse feature vectors; ii) pragmatic message representation, which achieves pragmatic approximation of high-dimensional feature vectors with a task-adaptive dictionary, enabling communicating with integer indices; iii) pragmatic collaborator selection, which identifies beneficial collaborators, pruning unnecessary communication links. Following this strategy, we first formulate a mathematical optimization framework for the perception-communication trade-off and then propose PragComm, a multi-agent collaborative perception system with two key components: i) single-agent detection and tracking and ii) pragmatic collaboration. The proposed PragComm promotes pragmatic communication and adapts to a wide range of communication conditions. We evaluate PragComm for both collaborative 3D object detection and tracking tasks in both real-world, V2V4Real, and simulation datasets, OPV2V and V2X-SIM2.0. PragComm consistently outperforms previous methods with more than 32.7K times lower communication volume on OPV2V. Code is available at github.com/PhyllisH/PragComm.
Enhancing Object Detection Performance for Small Objects through Synthetic Data Generation and Proportional Class-Balancing Technique: A Comparative Study in Industrial Scenarios
Authors: Authors: Jibinraj Antony, Vinit Hegiste, Ali Nazeri, Hooman Tavakoli, Snehal Walunj, Christiane Plociennik, Martin Ruskowski
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Abstract
Object Detection (OD) has proven to be a significant computer vision method in extracting localized class information and has multiple applications in the industry. Although many of the state-of-the-art (SOTA) OD models perform well on medium and large sized objects, they seem to under perform on small objects. In most of the industrial use cases, it is difficult to collect and annotate data for small objects, as it is time-consuming and prone to human errors. Additionally, those datasets are likely to be unbalanced and often result in an inefficient model convergence. To tackle this challenge, this study presents a novel approach that injects additional data points to improve the performance of the OD models. Using synthetic data generation, the difficulties in data collection and annotations for small object data points can be minimized and to create a dataset with balanced distribution. This paper discusses the effects of a simple proportional class-balancing technique, to enable better anchor matching of the OD models. A comparison was carried out on the performances of the SOTA OD models: YOLOv5, YOLOv7 and SSD, for combinations of real and synthetic datasets within an industrial use case.
MORPH: Towards Automated Concept Drift Adaptation for Malware Detection
Abstract
Concept drift is a significant challenge for malware detection, as the performance of trained machine learning models degrades over time, rendering them impractical. While prior research in malware concept drift adaptation has primarily focused on active learning, which involves selecting representative samples to update the model, self-training has emerged as a promising approach to mitigate concept drift. Self-training involves retraining the model using pseudo labels to adapt to shifting data distributions. In this research, we propose MORPH -- an effective pseudo-label-based concept drift adaptation method specifically designed for neural networks. Through extensive experimental analysis of Android and Windows malware datasets, we demonstrate the efficacy of our approach in mitigating the impact of concept drift. Our method offers the advantage of reducing annotation efforts when combined with active learning. Furthermore, our method significantly improves over existing works in automated concept drift adaptation for malware detection.
Deep Learning-based Target-To-User Association in Integrated Sensing and Communication Systems
Abstract
In Integrated Sensing and Communication (ISAC) systems, matching the radar targets with communication user equipments (UEs) is functional to several communication tasks, such as proactive handover and beam prediction. In this paper, we consider a radar-assisted communication system where a base station (BS) is equipped with a multiple-input-multiple-output (MIMO) radar that has a double aim: (i) associate vehicular radar targets to vehicular equipments (VEs) in the communication beamspace and (ii) predict the beamforming vector for each VE from radar data. The proposed target-to-user (T2U) association consists of two stages. First, vehicular radar targets are detected from range-angle images, and, for each, a beamforming vector is estimated. Then, the inferred per-target beamforming vectors are matched with the ones utilized at the BS for communication to perform target-to-user (T2U) association. Joint multi-target detection and beam inference is obtained by modifying the you only look once (YOLO) model, which is trained over simulated range-angle radar images. Simulation results over different urban vehicular mobility scenarios show that the proposed T2U method provides a probability of correct association that increases with the size of the BS antenna array, highlighting the respective increase of the separability of the VEs in the beamspace. Moreover, we show that the modified YOLO architecture can effectively perform both beam prediction and radar target detection, with similar performance in mean average precision on the latter over different antenna array sizes.
Unlocking the Potential: Multi-task Deep Learning for Spaceborne Quantitative Monitoring of Fugitive Methane Plumes
Authors: Authors: Guoxin Si, Shiliang Fu, Wei Yao
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
With the intensification of global warming, the monitoring of methane emission and detection of gas plumes from landfills have increasingly received attention. We decompose methane emission monitoring into three sub-tasks: methane concentration inversion, plume segmentation, and emission rate estimation. Conventional algorithms have limitations: methane concentration inversion usually uses the matched filter, which is sensitive to global spectrum distribution and contains a large amount of noises. There is limited research on plume segmentation, with many studies resorting to manual segmentation that is likely to be subjective. The estimation of methane emission rate often utilizes IME algorithm, which relies on obtaining meteorological measurement data. Using the WENT landfill site in Hong Kong and PRISMA hyperspectral satellite imagery, we propose a new deep learning-based framework for quantitative monitoring of methane emissions from remote sensing images based on physical simulation. We generate simulated methane plumes using large eddy simulation (LES) and different concentration maps of fugitive emission using the radiative transfer equation (RTE), while combining augmentation techniques to create a simulated PRISMA dataset. We train a U-Net network for methane concentration inversion, a Mask R-CNN network for methane plume segmentation, and a ResNet-50 network for methane emission rate estimation. All three deep networks achieve higher validation accuracy compared to conventional algorithms. We further respectively combine the first two sub-tasks and the last two sub-tasks to design the multi-task learning models - MTL-01 and MTL-02, both of which achieve higher accuracy than single-task models. Our research serves as a demonstration of applying multi-task deep learning to quantitative methane monitoring and can be extended to a broad range of methane monitoring tasks.
FocusFlow: 3D Gaze-Depth Interaction in Virtual Reality Leveraging Active Visual Depth Manipulation
Authors: Authors: Chenyang Zhang, Tiansu Chen, Eric Shaffer, Elahe Soltanaghai
Abstract
Gaze interaction presents a promising avenue in Virtual Reality (VR) due to its intuitive and efficient user experience. Yet, the depth control inherent in our visual system remains underutilized in current methods. In this study, we introduce FocusFlow, a hands-free interaction method that capitalizes on human visual depth perception within the 3D scenes of Virtual Reality. We first develop a binocular visual depth detection algorithm to understand eye input characteristics. We then propose a layer-based user interface and introduce the concept of 'Virtual Window' that offers an intuitive and robust gaze-depth VR interaction, despite the constraints of visual depth accuracy and precision spatially at further distances. Finally, to help novice users actively manipulate their visual depth, we propose two learning strategies that use different visual cues to help users master visual depth control. Our user studies on 24 participants demonstrate the usability of our proposed virtual window concept as a gaze-depth interaction method. In addition, our findings reveal that the user experience can be enhanced through an effective learning process with adaptive visual cues, helping users to develop muscle memory for this brand-new input mechanism. We conclude the paper by discussing strategies to optimize learning and potential research topics of gaze-depth interaction.
Raidar: geneRative AI Detection viA Rewriting
Authors: Authors: Chengzhi Mao, Carl Vondrick, Hao Wang, Junfeng Yang
Abstract
We find that large language models (LLMs) are more likely to modify human-written text than AI-generated text when tasked with rewriting. This tendency arises because LLMs often perceive AI-generated text as high-quality, leading to fewer modifications. We introduce a method to detect AI-generated content by prompting LLMs to rewrite text and calculating the editing distance of the output. We dubbed our geneRative AI Detection viA Rewriting method Raidar. Raidar significantly improves the F1 detection scores of existing AI content detection models -- both academic and commercial -- across various domains, including News, creative writing, student essays, code, Yelp reviews, and arXiv papers, with gains of up to 29 points. Operating solely on word symbols without high-dimensional features, our method is compatible with black box LLMs, and is inherently robust on new content. Our results illustrate the unique imprint of machine-generated text through the lens of the machines themselves.
Keyword: face recognition
There is no result
Keyword: augmentation
Memorization in Self-Supervised Learning Improves Downstream Generalization
Authors: Authors: Wenhao Wang, Muhammad Ahmad Kaleem, Adam Dziedzic, Michael Backes, Nicolas Papernot, Franziska Boenisch
Abstract
Self-supervised learning (SSL) has recently received significant attention due to its ability to train high-performance encoders purely on unlabeled data-often scraped from the internet. This data can still be sensitive and empirical evidence suggests that SSL encoders memorize private information of their training data and can disclose them at inference time. Since existing theoretical definitions of memorization from supervised learning rely on labels, they do not transfer to SSL. To address this gap, we propose SSLMem, a framework for defining memorization within SSL. Our definition compares the difference in alignment of representations for data points and their augmented views returned by both encoders that were trained on these data points and encoders that were not. Through comprehensive empirical analysis on diverse encoder architectures and datasets we highlight that even though SSL relies on large datasets and strong augmentations-both known in supervised learning as regularization techniques that reduce overfitting-still significant fractions of training data points experience high memorization. Through our empirical results, we show that this memorization is essential for encoders to achieve higher generalization performance on different downstream tasks.
OCT-SelfNet: A Self-Supervised Framework with Multi-Modal Datasets for Generalized and Robust Retinal Disease Detection
Authors: Authors: Fatema-E Jannat, Sina Gholami, Minhaj Nur Alam, Hamed Tabkhi
Abstract
Despite the revolutionary impact of AI and the development of locally trained algorithms, achieving widespread generalized learning from multi-modal data in medical AI remains a significant challenge. This gap hinders the practical deployment of scalable medical AI solutions. Addressing this challenge, our research contributes a self-supervised robust machine learning framework, OCT-SelfNet, for detecting eye diseases using optical coherence tomography (OCT) images. In this work, various data sets from various institutions are combined enabling a more comprehensive range of representation. Our method addresses the issue using a two-phase training approach that combines self-supervised pretraining and supervised fine-tuning with a mask autoencoder based on the SwinV2 backbone by providing a solution for real-world clinical deployment. Extensive experiments on three datasets with different encoder backbones, low data settings, unseen data settings, and the effect of augmentation show that our method outperforms the baseline model, Resnet-50 by consistently attaining AUC-ROC performance surpassing 77% across all tests, whereas the baseline model exceeds 54%. Moreover, in terms of the AUC-PR metric, our proposed method exceeded 42%, showcasing a substantial increase of at least 10% in performance compared to the baseline, which exceeded only 33%. This contributes to our understanding of our approach's potential and emphasizes its usefulness in clinical settings.
NIV-SSD: Neighbor IoU-Voting Single-Stage Object Detector From Point Cloud
Authors: Authors: Shuai Liu, Di Wang, Quan Wang, Kai Huang
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
Previous single-stage detectors typically suffer the misalignment between localization accuracy and classification confidence. To solve the misalignment problem, we introduce a novel rectification method named neighbor IoU-voting (NIV) strategy. Typically, classification and regression are treated as separate branches, making it challenging to establish a connection between them. Consequently, the classification confidence cannot accurately reflect the regression quality. NIV strategy can serve as a bridge between classification and regression branches by calculating two types of statistical data from the regression output to correct the classification confidence. Furthermore, to alleviate the imbalance of detection accuracy for complete objects with dense points (easy objects) and incomplete objects with sparse points (difficult objects), we propose a new data augmentation scheme named object resampling. It undersamples easy objects and oversamples difficult objects by randomly transforming part of easy objects into difficult objects. Finally, combining the NIV strategy and object resampling augmentation, we design an efficient single-stage detector termed NIV-SSD. Extensive experiments on several datasets indicate the effectiveness of the NIV strategy and the competitive performance of the NIV-SSD detector. The code will be available at https://github.com/Say2L/NIV-SSD.
On Building Myopic MPC Policies using Supervised Learning
Authors: Authors: Christopher A. Orrico, Bokan Yang, Dinesh Krishnamoorthy
Subjects: Machine Learning (cs.LG); Systems and Control (eess.SY); Optimization and Control (math.OC)
Abstract
The application of supervised learning techniques in combination with model predictive control (MPC) has recently generated significant interest, particularly in the area of approximate explicit MPC, where function approximators like deep neural networks are used to learn the MPC policy via optimal state-action pairs generated offline. While the aim of approximate explicit MPC is to closely replicate the MPC policy, substituting online optimization with a trained neural network, the performance guarantees that come with solving the online optimization problem are typically lost. This paper considers an alternative strategy, where supervised learning is used to learn the optimal value function offline instead of learning the optimal policy. This can then be used as the cost-to-go function in a myopic MPC with a very short prediction horizon, such that the online computation burden reduces significantly without affecting the controller performance. This approach differs from existing work on value function approximations in the sense that it learns the cost-to-go function by using offline-collected state-value pairs, rather than closed-loop performance data. The cost of generating the state-value pairs used for training is addressed using a sensitivity-based data augmentation scheme.
Graph Contrastive Invariant Learning from the Causal Perspective
Authors: Authors: Yanhu Mo, Xiao Wang, Shaohua Fan, Chuan Shi
Subjects: Machine Learning (cs.LG); Social and Information Networks (cs.SI)
Abstract
Graph contrastive learning (GCL), learning the node representation by contrasting two augmented graphs in a self-supervised way, has attracted considerable attention. GCL is usually believed to learn the invariant representation. However, does this understanding always hold in practice? In this paper, we first study GCL from the perspective of causality. By analyzing GCL with the structural causal model (SCM), we discover that traditional GCL may not well learn the invariant representations due to the non-causal information contained in the graph. How can we fix it and encourage the current GCL to learn better invariant representations? The SCM offers two requirements and motives us to propose a novel GCL method. Particularly, we introduce the spectral graph augmentation to simulate the intervention upon non-causal factors. Then we design the invariance objective and independence objective to better capture the causal factors. Specifically, (i) the invariance objective encourages the encoder to capture the invariant information contained in causal variables, and (ii) the independence objective aims to reduce the influence of confounders on the causal variables. Experimental results demonstrate the effectiveness of our approach on node classification tasks.
Interpreting Equivariant Representations
Authors: Authors: Andreas Abildtrup Hansen, Anna Calissano, Aasa Feragen
Abstract
Latent representations are used extensively for downstream tasks, such as visualization, interpolation or feature extraction of deep learning models. Invariant and equivariant neural networks are powerful and well-established models for enforcing inductive biases. In this paper, we demonstrate that the inductive bias imposed on the by an equivariant model must also be taken into account when using latent representations. We show how not accounting for the inductive biases leads to decreased performance on downstream tasks, and vice versa, how accounting for inductive biases can be done effectively by using an invariant projection of the latent representations. We propose principles for how to choose such a projection, and show the impact of using these principles in two common examples: First, we study a permutation equivariant variational auto-encoder trained for molecule graph generation; here we show that invariant projections can be designed that incur no loss of information in the resulting invariant representation. Next, we study a rotation-equivariant representation used for image classification. Here, we illustrate how random invariant projections can be used to obtain an invariant representation with a high degree of retained information. In both cases, the analysis of invariant latent representations proves superior to their equivariant counterparts. Finally, we illustrate that the phenomena documented here for equivariant neural networks have counterparts in standard neural networks where invariance is encouraged via augmentation. Thus, while these ambiguities may be known by experienced developers of equivariant models, we make both the knowledge as well as effective tools to handle the ambiguities available to the broader community.
Context Matters: Pushing the Boundaries of Open-Ended Answer Generation with Graph-Structured Knowledge Context
Abstract
In the continuously advancing AI landscape, crafting context-rich and meaningful responses via Large Language Models (LLMs) is essential. Researchers are becoming more aware of the challenges that LLMs with fewer parameters encounter when trying to provide suitable answers to open-ended questions. To address these hurdles, the integration of cutting-edge strategies, augmentation of rich external domain knowledge to LLMs, offers significant improvements. This paper introduces a novel framework that combines graph-driven context retrieval in conjunction to knowledge graphs based enhancement, honing the proficiency of LLMs, especially in domain specific community question answering platforms like AskUbuntu, Unix, and ServerFault. We conduct experiments on various LLMs with different parameter sizes to evaluate their ability to ground knowledge and determine factual accuracy in answers to open-ended questions. Our methodology GraphContextGen consistently outperforms dominant text-based retrieval systems, demonstrating its robustness and adaptability to a larger number of use cases. This advancement highlights the importance of pairing context rich data retrieval with LLMs, offering a renewed approach to knowledge sourcing and generation in AI systems. We also show that, due to rich contextual data retrieval, the crucial entities, along with the generated answer, remain factually coherent with the gold answer.
Non-Neighbors Also Matter to Kriging: A New Contrastive-Prototypical Learning
Authors: Authors: Zhishuai Li, Yunhao Nie, Ziyue Li, Lei Bai, Yisheng Lv, Rui Zhao
Abstract
Kriging aims at estimating the attributes of unsampled geo-locations from observations in the spatial vicinity or physical connections, which helps mitigate skewed monitoring caused by under-deployed sensors. Existing works assume that neighbors' information offers the basis for estimating the attributes of the unobserved target while ignoring non-neighbors. However, non-neighbors could also offer constructive information, and neighbors could also be misleading. To this end, we propose ``Contrastive-Prototypical'' self-supervised learning for Kriging (KCP) to refine valuable information from neighbors and recycle the one from non-neighbors. As a pre-trained paradigm, we conduct the Kriging task from a new perspective of representation: we aim to first learn robust and general representations and then recover attributes from representations. A neighboring contrastive module is designed that coarsely learns the representations by narrowing the representation distance between the target and its neighbors while pushing away the non-neighbors. In parallel, a prototypical module is introduced to identify similar representations via exchanged prediction, thus refining the misleading neighbors and recycling the useful non-neighbors from the neighboring contrast component. As a result, not all the neighbors and some of the non-neighbors will be used to infer the target. To encourage the two modules above to learn general and robust representations, we design an adaptive augmentation module that incorporates data-driven attribute augmentation and centrality-based topology augmentation over the spatiotemporal Kriging graph data. Extensive experiments on real-world datasets demonstrate the superior performance of KCP compared to its peers with 6% improvements and exceptional transferability and robustness. The code is available at https://github.com/bonaldli/KCP
Unlocking the Potential: Multi-task Deep Learning for Spaceborne Quantitative Monitoring of Fugitive Methane Plumes
Authors: Authors: Guoxin Si, Shiliang Fu, Wei Yao
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Abstract
With the intensification of global warming, the monitoring of methane emission and detection of gas plumes from landfills have increasingly received attention. We decompose methane emission monitoring into three sub-tasks: methane concentration inversion, plume segmentation, and emission rate estimation. Conventional algorithms have limitations: methane concentration inversion usually uses the matched filter, which is sensitive to global spectrum distribution and contains a large amount of noises. There is limited research on plume segmentation, with many studies resorting to manual segmentation that is likely to be subjective. The estimation of methane emission rate often utilizes IME algorithm, which relies on obtaining meteorological measurement data. Using the WENT landfill site in Hong Kong and PRISMA hyperspectral satellite imagery, we propose a new deep learning-based framework for quantitative monitoring of methane emissions from remote sensing images based on physical simulation. We generate simulated methane plumes using large eddy simulation (LES) and different concentration maps of fugitive emission using the radiative transfer equation (RTE), while combining augmentation techniques to create a simulated PRISMA dataset. We train a U-Net network for methane concentration inversion, a Mask R-CNN network for methane plume segmentation, and a ResNet-50 network for methane emission rate estimation. All three deep networks achieve higher validation accuracy compared to conventional algorithms. We further respectively combine the first two sub-tasks and the last two sub-tasks to design the multi-task learning models - MTL-01 and MTL-02, both of which achieve higher accuracy than single-task models. Our research serves as a demonstration of applying multi-task deep learning to quantitative methane monitoring and can be extended to a broad range of methane monitoring tasks.
Keyword: detection
Multimodal Data Curation via Object Detection and Filter Ensembles
A Lightweight FPGA-based IDS-ECU Architecture for Automotive CAN
Quantised Neural Network Accelerators for Low-Power IDS in Automotive Networks
Machine learning-based network intrusion detection for big and imbalanced data using oversampling, stacking feature embedding and feature extraction
Fine-tuning Large Language Models for Multigenerator, Multidomain, and Multilingual Machine-Generated Text Detection
ACS: Concurrent Kernel Execution on Irregular, Input-Dependent Computational Graphs
Learning Recovery Strategies for Dynamic Self-healing in Reactive Systems
Patch2QL: Discover Cognate Defects in Open Source Software Supply Chain With Auto-generated Static Analysis Rules
NIV-SSD: Neighbor IoU-Voting Single-Stage Object Detector From Point Cloud
Self-supervised Learning of LiDAR 3D Point Clouds via 2D-3D Neural Calibration
Comparing Human-Centered Language Modeling: Is it Better to Model Groups, Individual Traits, or Both?
On the Fundamental Tradeoff of Joint Communication and Quickest Change Detection
Small Language Model Meets with Reinforced Vision Vocabulary
Open-Set Facial Expression Recognition
Detecting and recognizing characters in Greek papyri with YOLOv8, DeiT and SimCLR
Multi-Party Private Set Intersection: A Circuit-Based Protocol with Jaccard Similarity for Secure and Efficient Anomaly Detection in Network Traffic
Blind Channel Estimation and Joint Symbol Detection with Data-Driven Factor Graphs
Pragmatic Communication in Multi-Agent Collaborative Perception
Enhancing Object Detection Performance for Small Objects through Synthetic Data Generation and Proportional Class-Balancing Technique: A Comparative Study in Industrial Scenarios
MORPH: Towards Automated Concept Drift Adaptation for Malware Detection
Deep Learning-based Target-To-User Association in Integrated Sensing and Communication Systems
Unlocking the Potential: Multi-task Deep Learning for Spaceborne Quantitative Monitoring of Fugitive Methane Plumes
FocusFlow: 3D Gaze-Depth Interaction in Virtual Reality Leveraging Active Visual Depth Manipulation
Raidar: geneRative AI Detection viA Rewriting
Keyword: face recognition
There is no result
Keyword: augmentation
Memorization in Self-Supervised Learning Improves Downstream Generalization
OCT-SelfNet: A Self-Supervised Framework with Multi-Modal Datasets for Generalized and Robust Retinal Disease Detection
NIV-SSD: Neighbor IoU-Voting Single-Stage Object Detector From Point Cloud
On Building Myopic MPC Policies using Supervised Learning
Graph Contrastive Invariant Learning from the Causal Perspective
Interpreting Equivariant Representations
Context Matters: Pushing the Boundaries of Open-Ended Answer Generation with Graph-Structured Knowledge Context
Non-Neighbors Also Matter to Kriging: A New Contrastive-Prototypical Learning
Unlocking the Potential: Multi-task Deep Learning for Spaceborne Quantitative Monitoring of Fugitive Methane Plumes