Abstract
Large Language Models (LLMs) have shown promising potentials in program generation and no-code automation. However, LLMs are prone to generate hallucinations, i.e., they generate text which sounds plausible but is incorrect. Although there has been a recent surge in research on LLM hallucinations for text generation, similar hallucination phenomenon can happen in code generation. Sometimes the generated code can have syntactical or logical errors as well as more advanced issues like security vulnerabilities, memory leaks, etc. Given the wide adaptation of LLMs to enhance efficiency in code generation and development in general, it becomes imperative to investigate hallucinations in code generation. To the best of our knowledge, this is the first attempt at studying hallucinations in the code generated by LLMs. We start by introducing the code hallucination definition and a comprehensive taxonomy of code hallucination types. We propose the first benchmark CodeMirage dataset for code hallucinations. The benchmark contains 1,137 GPT-3.5 generated hallucinated code snippets for Python programming problems from two base datasets - HumanEval and MBPP. We then propose the methodology for code hallucination detection and experiment with open source LLMs such as CodeLLaMA as well as OpenAI's GPT-3.5 and GPT-4 models using one-shot prompt. We find that GPT-4 performs the best on HumanEval dataset and gives comparable results to the fine-tuned CodeBERT baseline on MBPP dataset. Towards the end, we discuss various mitigation strategies for code hallucinations and conclude our work.
Title:
5%>100%: Breaking Performance Shackles of Full Fine-Tuning on Visual Recognition Tasks
Authors: Dongshuo Yin, Leiyi Hu, Bin Li, Youqun Zhang, Xue Yang
Subjects: Subjects:
Computer Vision and Pattern Recognition (cs.CV)
Abstract
Pre-training & fine-tuning can enhance the transferring efficiency and performance in visual tasks. Recent delta-tuning methods provide more options for visual classification tasks. Despite their success, existing visual delta-tuning art fails to exceed the upper limit of full fine-tuning on challenging tasks like object detection and segmentation. To find a competitive alternative to full fine-tuning, we propose the Multi-cognitive Visual Adapter (Mona) tuning, a novel adapter-based tuning method. First, we introduce multiple vision-friendly filters into the adapter to enhance its ability to process visual signals, while previous methods mainly rely on language-friendly linear filters. Second, we add the scaled normalization layer in the adapter to regulate the distribution of input features for visual filters. To fully demonstrate the practicality and generality of Mona, we conduct experiments on multiple representative visual tasks, including instance segmentation on COCO, semantic segmentation on ADE20K, object detection on Pascal VOC, oriented object detection on DOTA/STAR, and image classification on three common datasets. Exciting results illustrate that Mona surpasses full fine-tuning on all these tasks, and is the only delta-tuning method outperforming full fine-tuning on the above various tasks. For example, Mona achieves 1% performance gain on the COCO dataset compared to full fine-tuning. Comprehensive results suggest that Mona-tuning is more suitable for retaining and utilizing the capabilities of pre-trained models than full fine-tuning. We will make the code publicly available.
Title:
Penny-Wise and Pound-Foolish in Deepfake Detection
Authors: Yabin Wang, Zhiwu Huang, Su Zhou, Adam Prugel-Bennett, Xiaopeng Hong
Subjects: Subjects:
Computer Vision and Pattern Recognition (cs.CV)
Abstract
The diffusion of deepfake technologies has sparked serious concerns about its potential misuse across various domains, prompting the urgent need for robust detection methods. Despite advancement, many current approaches prioritize short-term gains at expense of long-term effectiveness. This paper critiques the overly specialized approach of fine-tuning pre-trained models solely with a penny-wise objective on a single deepfake dataset, while disregarding the pound-wise balance for generalization and knowledge retention. To address this "Penny-Wise and Pound-Foolish" issue, we propose a novel learning framework (PoundNet) for generalization of deepfake detection on a pre-trained vision-language model. PoundNet incorporates a learnable prompt design and a balanced objective to preserve broad knowledge from upstream tasks (object classification) while enhancing generalization for downstream tasks (deepfake detection). We train PoundNet on a standard single deepfake dataset, following common practice in the literature. We then evaluate its performance across 10 public large-scale deepfake datasets with 5 main evaluation metrics-forming the largest benchmark test set for assessing the generalization ability of deepfake detection models, to our knowledge. The comprehensive benchmark evaluation demonstrates the proposed PoundNet is significantly less "Penny-Wise and Pound-Foolish", achieving a remarkable improvement of 19% in deepfake detection performance compared to state-of-the-art methods, while maintaining a strong performance of 63% on object classification tasks, where other deepfake detection models tend to be ineffective. Code and data are open-sourced at this https URL.
Title:
A Robust Multi-Stage Intrusion Detection System for In-Vehicle Network Security using Hierarchical Federated Learning
Authors: Muzun Althunayyan, Amir Javed, Omer Rana
Subjects: Subjects:
Cryptography and Security (cs.CR)
Abstract
As connected and autonomous vehicles proliferate, the Controller Area Network (CAN) bus has become the predominant communication standard for in-vehicle networks due to its speed and efficiency. However, the CAN bus lacks basic security measures such as authentication and encryption, making it highly vulnerable to cyberattacks. To ensure in-vehicle security, intrusion detection systems (IDSs) must detect seen attacks and provide a robust defense against new, unseen attacks while remaining lightweight for practical deployment. Previous work has relied solely on the CAN ID feature or has used traditional machine learning (ML) approaches with manual feature extraction. These approaches overlook other exploitable features, making it challenging to adapt to new unseen attack variants and compromising security. This paper introduces a cutting-edge, novel, lightweight, in-vehicle, IDS-leveraging, deep learning (DL) algorithm to address these limitations. The proposed IDS employs a multi-stage approach: an artificial neural network (ANN) in the first stage to detect seen attacks, and a Long Short-Term Memory (LSTM) autoencoder in the second stage to detect new, unseen attacks. To understand and analyze diverse driving behaviors, update the model with the latest attack patterns, and preserve data privacy, we propose a theoretical framework to deploy our IDS in a hierarchical federated learning (H-FL) environment. Experimental results demonstrate that our IDS achieves an F1-score exceeding 0.99 for seen attacks and exceeding 0.95 for novel attacks, with a detection rate of 99.99%. Additionally, the false alarm rate (FAR) is exceptionally low at 0.016%, minimizing false alarms. Despite using DL algorithms known for their effectiveness in identifying sophisticated and zero-day attacks, the IDS remains lightweight, ensuring its feasibility for real-world deployment.
Abstract
Traditional shadow detectors often identify all shadow regions of static images or video sequences. This work presents the Referring Video Shadow Detection (RVSD), which is an innovative task that rejuvenates the classic paradigm by facilitating the segmentation of particular shadows in videos based on descriptive natural language prompts. This novel RVSD not only achieves segmentation of arbitrary shadow areas of interest based on descriptions (flexibility) but also allows users to interact with visual content more directly and naturally by using natural language prompts (interactivity), paving the way for abundant applications ranging from advanced video editing to virtual reality experiences. To pioneer the RVSD research, we curated a well-annotated RVSD dataset, which encompasses 86 videos and a rich set of 15,011 paired textual descriptions with corresponding shadows. To the best of our knowledge, this dataset is the first one for addressing RVSD. Based on this dataset, we propose a Referring Shadow-Track Memory Network (RSM-Net) for addressing the RVSD task. In our RSM-Net, we devise a Twin-Track Synergistic Memory (TSM) to store intra-clip memory features and hierarchical inter-clip memory features, and then pass these memory features into a memory read module to refine features of the current video frame for referring shadow detection. We also develop a Mixed-Prior Shadow Attention (MSA) to utilize physical priors to obtain a coarse shadow map for learning more visual features by weighting it with the input video frame. Experimental results show that our RSM-Net achieves state-of-the-art performance for RVSD with a notable Overall IOU increase of 4.4\%. Our code and dataset are available at this https URL.
Title:
Vulnerability Handling of AI-Generated Code -- Existing Solutions and Open Challenges
Abstract
The increasing use of generative Artificial Intelligence (AI) in modern software engineering, particularly Large Language Models (LLMs) for code generation, has transformed professional software development by boosting productivity and automating development processes. This adoption, however, has highlighted a significant issue: the introduction of security vulnerabilities into the code. These vulnerabilities result, e.g., from flaws in the training data that propagate into the generated code, creating challenges in disclosing them. Traditional vulnerability handling processes often involve extensive manual review. Applying such traditional processes to AI-generated code is challenging. AI-generated code may include several vulnerabilities, possibly in slightly different forms as developers might not build on already implemented code but prompt similar tasks. In this work, we explore the current state of LLM-based approaches for vulnerability handling, focusing on approaches for vulnerability detection, localization, and repair. We provide an overview of recent progress in this area and highlight open challenges that must be addressed in order to establish a reliable and scalable vulnerability handling process of AI-generated code.
Title:
Integrating Multi-view Analysis: Multi-view Mixture-of-Expert for Textual Personality Detection
Abstract
Textual personality detection aims to identify personality traits by analyzing user-generated content. To achieve this effectively, it is essential to thoroughly examine user-generated content from various perspectives. However, previous studies have struggled with automatically extracting and effectively integrating information from multiple perspectives, thereby limiting their performance on personality detection. To address these challenges, we propose the Multi-view Mixture-of-Experts Model for Textual Personality Detection (MvP). MvP introduces a Multi-view Mixture-of-Experts (MoE) network to automatically analyze user posts from various perspectives. Additionally, it employs User Consistency Regularization to mitigate conflicts among different perspectives and learn a multi-view generic user representation. The model's training is optimized via a multi-task joint learning strategy that balances supervised personality detection with self-supervised user consistency constraints. Experimental results on two widely-used personality detection datasets demonstrate the effectiveness of the MvP model and the benefits of automatically analyzing user posts from diverse perspectives for textual personality detection.
Title:
Enhancing Discriminative Tasks by Guiding the Pre-trained Language Model with Large Language Model's Experience
Abstract
Large Language Models (LLMs) and pre-trained Language Models (LMs) have achieved impressive success on many software engineering tasks (e.g., code completion and code generation). By leveraging huge existing code corpora (e.g., GitHub), these models aim to understand the patterns in source code and use these patterns to predict code properties. However, fine-tuning LLMs is time-consuming and costly for end users and small organizations. Furthermore, fine-tuning LMs heavily depends on the amount and quality of datasets available. As a result, the current lack of data and the high cost of collecting it in real-world scenarios further limit the applicability of LMs. In this paper, we leverage the powerful generation capabilities of LLMs to enhance pre-trained LMs. Specifically, we use LLMs to generate domain-specific data, thereby improving the performance of pre-trained LMs on the target tasks. We conduct experiments by combining different LLMs in our generation phase and introducing various LMs to learn from the LLM-generated data. Then, we compare the performance of these LMs before and after learning the data. We find that LLM-generated data significantly enhances the performance of LMs. The improvement can reach up to 58.36% for fault localization and up to 6.09% for clone detection. Our study highlights that using LLMs to generate data for LMs can improve performance by a large margin.
Title:
Detection and tracking of MAVs using a LiDAR with rosette scanning pattern
Authors: Sándor Gazdag, Tom Möller, Tamás Filep, Anita Keszler, András L. Majdik
Subjects: Subjects:
Robotics (cs.RO); Computer Vision and Pattern Recognition (cs.CV)
Abstract
The usage of commercial Micro Aerial Vehicles (MAVs) has increased drastically during the last decade. While the added value of MAVs to society is apparent, their growing use is also coming with increasing risks like violating public airspace at airports or committing privacy violations. To mitigate these issues it is becoming critical to develop solutions that incorporate the detection and tracking of MAVs with autonomous systems. This work presents a method for the detection and tracking of MAVs using a novel, low-cost rosette scanning LiDAR on a pan-tilt turret. Once the static background is captured, a particle filter is utilized to detect a possible target and track its position with a physical, programmable pan-tilt system. The tracking makes it possible to keep the MAV in the center, maximizing the density of 3D points measured on the target by the LiDAR sensor. The developed algorithm was evaluated within the indoor MIcro aerial vehicle and MOtion capture (MIMO) arena and has state-of-the-art tracking accuracy, stability, and fast re-detection time in case of tracking loss. Based on the outdoor tests, it was possible to significantly increase the detection distance and number of returned points compared to other similar methods using LiDAR.
Title:
A training regime to learn unified representations from complementary breast imaging modalities
Authors: Umang Sharma, Jungkyu Park, Laura Heacock, Sumit Chopra, Krzysztof Geras
Abstract
Full Field Digital Mammograms (FFDMs) and Digital Breast Tomosynthesis (DBT) are the two most widely used imaging modalities for breast cancer screening. Although DBT has increased cancer detection compared to FFDM, its widespread adoption in clinical practice has been slowed by increased interpretation times and a perceived decrease in the conspicuity of specific lesion types. Specifically, the non-inferiority of DBT for microcalcifications remains under debate. Due to concerns about the decrease in visual acuity, combined DBT-FFDM acquisitions remain popular, leading to overall increased exam times and radiation dosage. Enabling DBT to provide diagnostic information present in both FFDM and DBT would reduce reliance on FFDM, resulting in a reduction in both quantities. We propose a machine learning methodology that learns high-level representations leveraging the complementary diagnostic signal from both DBT and FFDM. Experiments on a large-scale data set validate our claims and show that our representations enable more accurate breast lesion detection than any DBT- or FFDM-based model.
Title:
Tell Codec What Worth Compressing: Semantically Disentangled Image Coding for Machine with LMMs
Abstract
We present a new image compression paradigm to achieve intelligently coding for machine'' by cleverly leveraging the common sense of Large Multimodal Models (LMMs). We are motivated by the evidence that large language/multimodal models are powerful general-purpose semantics predictors for understanding the real world. Different from traditional image compression typically optimized for human eyes, the image coding for machines (ICM) framework we focus on requires the compressed bitstream to more comply with different downstream intelligent analysis tasks. To this end, we employ LMM to \textcolor{red}{tell codec what to compress}: 1) first utilize the powerful semantic understanding capability of LMMs w.r.t object grounding, identification, and importance ranking via prompts, to disentangle image content before compression, 2) and then based on these semantic priors we accordingly encode and transmit objects of the image in order with a structured bitstream. In this way, diverse vision benchmarks including image classification, object detection, instance segmentation, etc., can be well supported with such a semantically structured bitstream. We dub our method\textit{SDComp}'' for ``\textit{S}emantically \textit{D}isentangled \textit{Comp}ression'', and compare it with state-of-the-art codecs on a wide variety of different vision tasks. SDComp codec leads to more flexible reconstruction results, promised decoded visual quality, and a more generic/satisfactory intelligent task-supporting ability.
Title:
DeepREST: Automated Test Case Generation for REST APIs Exploiting Deep Reinforcement Learning
Abstract
Automatically crafting test scenarios for REST APIs helps deliver more reliable and trustworthy web-oriented systems. However, current black-box testing approaches rely heavily on the information available in the API's formal documentation, i.e., the OpenAPI Specification (OAS for short). While useful, the OAS mostly covers syntactic aspects of the API (e.g., producer-consumer relations between operations, input value properties, and additional constraints in natural language), and it lacks a deeper understanding of the API business logic. Missing semantics include implicit ordering (logic dependency) between operations and implicit input-value constraints. These limitations hinder the ability of black-box testing tools to generate truly effective test cases automatically. This paper introduces DeepREST, a novel black-box approach for automatically testing REST APIs. It leverages deep reinforcement learning to uncover implicit API constraints, that is, constraints hidden from API documentation. Curiosity-driven learning guides an agent in the exploration of the API and learns an effective order to test its operations. This helps identify which operations to test first to take the API in a testable state and avoid failing API interactions later. At the same time, experience gained on successful API interactions is leveraged to drive accurate input data generation (i.e., what parameters to use and how to pick their values). Additionally, DeepREST alternates exploration with exploitation by mutating successful API interactions to improve test coverage and collect further experience. Our empirical validation suggests that the proposed approach is very effective in achieving high test coverage and fault detection and superior to a state-of-the-art baseline.
Title:
Enhanced BICM Receiver Metrics for Ultra-Reliable Low-Latency Short Block Channel Communications
Abstract
This paper presents Bit-Interleaved coded modulation metrics for joint estimation and detection in short block length channels, addressing scenarios with unknown channel state information and low training dimension density. We show that it is possible to enhance the performance and sensitivity through joint detection-estimation compared to standard receivers, especially when the channel state information is unknown and the density of the training dimensions is low. The performance analysis makes use of a full 5G transmitter and receiver chains for both Polar and LDPC coded transmissions paired with QPSK modulation scheme. We consider transmissions where reference signals are interleaved with data and both are transmitted over a small number of OFDM symbols so that near-perfect channel estimation cannot be achieved. This is particularly adapted to mini-slot transmissions for ultra-reliable, low-latency communications (URLLC) or for short packet random access use cases. Performance evaluation spans SIMO and SU-MIMO configurations, emphasizing the efficacy of BICM detection in realistic base station receiver scenarios. Our findings demonstrate that when the detection windows used in the metric units are on the order of four modulated symbols, the proposed receiver metrics can be used to achieve detection performance that is close to that of a coherent receiver with perfect channel state information.
Title:
MIA-Tuner: Adapting Large Language Models as Pre-training Text Detector
Abstract
The increasing parameters and expansive dataset of large language models (LLMs) highlight the urgent demand for a technical solution to audit the underlying privacy risks and copyright issues associated with LLMs. Existing studies have partially addressed this need through an exploration of the pre-training data detection problem, which is an instance of a membership inference attack (MIA). This problem involves determining whether a given piece of text has been used during the pre-training phase of the target LLM. Although existing methods have designed various sophisticated MIA score functions to achieve considerable detection performance in pre-trained LLMs, how to achieve high-confidence detection and how to perform MIA on aligned LLMs remain challenging. In this paper, we propose MIA-Tuner, a novel instruction-based MIA method, which instructs LLMs themselves to serve as a more precise pre-training data detector internally, rather than design an external MIA score function. Furthermore, we design two instruction-based safeguards to respectively mitigate the privacy risks brought by the existing methods and MIA-Tuner. To comprehensively evaluate the most recent state-of-the-art LLMs, we collect a more up-to-date MIA benchmark dataset, named WIKIMIA-24, to replace the widely adopted benchmark WIKIMIA. We conduct extensive experiments across various aligned and unaligned LLMs over the two benchmark datasets. The results demonstrate that MIA-Tuner increases the AUC of MIAs from 0.7 to a significantly high level of 0.9.
Title:
MAT-SED: AMasked Audio Transformer with Masked-Reconstruction Based Pre-training for Sound Event Detection
Authors: Pengfei Cai, Yan Song, Kang Li, Haoyu Song, Ian McLoughlin
Abstract
Sound event detection (SED) methods that leverage a large pre-trained Transformer encoder network have shown promising performance in recent DCASE challenges. However, they still rely on an RNN-based context network to model temporal dependencies, largely due to the scarcity of labeled data. In this work, we propose a pure Transformer-based SED model with masked-reconstruction based pre-training, termed MAT-SED. Specifically, a Transformer with relative positional encoding is first designed as the context network, pre-trained by the masked-reconstruction task on all available target data in a self-supervised way. Both the encoder and the context network are jointly fine-tuned in a semi-supervised manner. Furthermore, a global-local feature fusion strategy is proposed to enhance the localization capability. Evaluation of MAT-SED on DCASE2023 task4 surpasses state-of-the-art performance, achieving 0.587/0.896 PSDS1/PSDS2 respectively.
Title:
Multimodal Relational Triple Extraction with Query-based Entity Object Transformer
Authors: Lei Hei, Ning An, Tingjing Liao, Qi Ma, Jiaqi Wang, Feiliang Ren
Abstract
Multimodal Relation Extraction is crucial for constructing flexible and realistic knowledge graphs. Recent studies focus on extracting the relation type with entity pairs present in different modalities, such as one entity in the text and another in the image. However, existing approaches require entities and objects given beforehand, which is costly and impractical. To address the limitation, we propose a novel task, Multimodal Entity-Object Relational Triple Extraction, which aims to extract all triples (entity span, relation, object region) from image-text pairs. To facilitate this study, we modified a multimodal relation extraction dataset MORE, which includes 21 relation types, to create a new dataset containing 20,264 triples, averaging 5.75 triples per image-text pair. Moreover, we propose QEOT, a query-based model with a selective attention mechanism, to dynamically explore the interaction and fusion of textual and visual information. In particular, the proposed method can simultaneously accomplish entity extraction, relation classification, and object detection with a set of queries. Our method is suitable for downstream applications and reduces error accumulation due to the pipeline-style approaches. Extensive experimental results demonstrate that our proposed method outperforms the existing baselines by 8.06% and achieves state-of-the-art performance.
Title:
A Novel Buffered Federated Learning Framework for Privacy-Driven Anomaly Detection in IIoT
Abstract
Industrial Internet of Things (IIoT) is highly sensitive to data privacy and cybersecurity threats. Federated Learning (FL) has emerged as a solution for preserving privacy, enabling private data to remain on local IIoT clients while cooperatively training models to detect network anomalies. However, both synchronous and asynchronous FL architectures exhibit limitations, particularly when dealing with clients with varying speeds due to data heterogeneity and resource constraints. Synchronous architecture suffers from straggler effects, while asynchronous methods encounter communication bottlenecks. Additionally, FL models are prone to adversarial inference attacks aimed at disclosing private training data. To address these challenges, we propose a Buffered FL (BFL) framework empowered by homomorphic encryption for anomaly detection in heterogeneous IIoT environments. BFL utilizes a novel weighted average time approach to mitigate both straggler effects and communication bottlenecks, ensuring fairness between clients with varying processing speeds through collaboration with a buffer-based server. The performance results, derived from two datasets, show the superiority of BFL compared to state-of-the-art FL methods, demonstrating improved accuracy and convergence speed while enhancing privacy preservation.
Title:
Xpikeformer: Hybrid Analog-Digital Hardware Acceleration for Spiking Transformers
Abstract
This paper introduces Xpikeformer, a hybrid analog-digital hardware architecture designed to accelerate spiking neural network (SNN)-based transformer models. By combining the energy efficiency and temporal dynamics of SNNs with the powerful sequence modeling capabilities of transformers, Xpikeformer leverages mixed analog-digital computing techniques to enhance performance and energy efficiency. The architecture integrates analog in-memory computing (AIMC) for feedforward and fully connected layers, and a stochastic spiking attention (SSA) engine for efficient attention mechanisms. We detail the design, implementation, and evaluation of Xpikeformer, demonstrating significant improvements in energy consumption and computational efficiency. Through an image classification task and a wireless communication symbol detection task, we show that Xpikeformer can achieve software-comparable inference accuracy. Energy evaluations reveal that Xpikeformer achieves up to a $17.8$--$19.2\times$ reduction in energy consumption compared to state-of-the-art digital ANN transformers and up to a $5.9$--$6.8\times$ reduction compared to fully digital SNN transformers. Xpikeformer also achieves a $12.0\times$ speedup compared to the GPU implementation of spiking transformers.
Abstract
Backscatter communication offers a promising solution to connect massive Internet-of-Things (IoT) devices with low cost and high energy efficiency. Nevertheless, its inherently passive nature limits transmission reliability, thereby hindering improvements in communication range and data rate. To overcome these challenges, we introduce a bistatic broadband backscatter communication (BBBC) system, which equips the backscatter device (BD) with multiple antennas. In the proposed BBBC system, a radio frequency (RF) source directs a sinusoidal signal to the BD, facilitating single-carrier block transmission at the BD. Meanwhile, without requiring channel state information (CSI), cyclic delay diversity (CDD) is employed at the multi-antenna BD to enhance transmission reliability through additional cyclically delayed backscattered signals. We also propose a receiver design that includes preprocessing of the time-domain received signal, pilot-based parameter estimation, and frequency-domain equalization, enabling low-complexity detection of the backscattered signal. Leveraging the matched filter bound (MFB), we analyze the achievable diversity gains in terms of outage probability. Our analysis reveals that spatial diversity is achievable under general Rayleigh fading conditions, and both frequency and spatial diversity are attainable in scenarios where the forward link experiences a line-of-sight (LoS) channel. Simulation results validate the effectiveness of the proposed BBBC system. As the number of BD antennas increases, our results show that the proposed scheme not only enhances array gain but also improves diversity order, significantly reducing both outage probability and bit error rate (BER). Consequently, it outperforms conventional schemes that yield only minor gains.
Title:
SAM2-UNet: Segment Anything 2 Makes Strong Encoder for Natural and Medical Image Segmentation
Authors: Xinyu Xiong, Zihuang Wu, Shuangyi Tan, Wenxue Li, Feilong Tang, Ying Chen, Siying Li, Jie Ma, Guanbin Li
Subjects: Subjects:
Computer Vision and Pattern Recognition (cs.CV)
Abstract
Image segmentation plays an important role in vision understanding. Recently, the emerging vision foundation models continuously achieved superior performance on various tasks. Following such success, in this paper, we prove that the Segment Anything Model 2 (SAM2) can be a strong encoder for U-shaped segmentation models. We propose a simple but effective framework, termed SAM2-UNet, for versatile image segmentation. Specifically, SAM2-UNet adopts the Hiera backbone of SAM2 as the encoder, while the decoder uses the classic U-shaped design. Additionally, adapters are inserted into the encoder to allow parameter-efficient fine-tuning. Preliminary experiments on various downstream tasks, such as camouflaged object detection, salient object detection, marine animal segmentation, mirror detection, and polyp segmentation, demonstrate that our SAM2-UNet can simply beat existing specialized state-of-the-art methods without bells and whistles. Project page: \url{this https URL}.
Keyword: face recognition
There is no result
Keyword: augmentation
Title:
Assessing and Enhancing Large Language Models in Rare Disease Question-answering
Abstract
Despite the impressive capabilities of Large Language Models (LLMs) in general medical domains, questions remain about their performance in diagnosing rare diseases. To answer this question, we aim to assess the diagnostic performance of LLMs in rare diseases, and explore methods to enhance their effectiveness in this area. In this work, we introduce a rare disease question-answering (ReDis-QA) dataset to evaluate the performance of LLMs in diagnosing rare diseases. Specifically, we collected 1360 high-quality question-answer pairs within the ReDis-QA dataset, covering 205 rare diseases. Additionally, we annotated meta-data for each question, facilitating the extraction of subsets specific to any given disease and its property. Based on the ReDis-QA dataset, we benchmarked several open-source LLMs, revealing that diagnosing rare diseases remains a significant challenge for these models. To facilitate retrieval augmentation generation for rare disease diagnosis, we collect the first rare diseases corpus (ReCOP), sourced from the National Organization for Rare Disorders (NORD) database. Specifically, we split the report of each rare disease into multiple chunks, each representing a different property of the disease, including their overview, symptoms, causes, effects, related disorders, diagnosis, and standard therapies. This structure ensures that the information within each chunk aligns consistently with a question. Experiment results demonstrate that ReCOP can effectively improve the accuracy of LLMs on the ReDis-QA dataset by an average of 8%. Moreover, it significantly guides LLMs to generate trustworthy answers and explanations that can be traced back to existing literature.
Title:
A semi-centralized multi-agent RL framework for efficient irrigation scheduling
Authors: Bernard T. Agyeman, Benjamin Decard-Nelson, Jinfeng Liu (University of Alberta), Sirish L. Shah
Subjects: Subjects:
Systems and Control (eess.SY); Machine Learning (cs.LG)
Abstract
This paper proposes a Semi-Centralized Multi-Agent Reinforcement Learning (SCMARL) approach for irrigation scheduling in spatially variable agricultural fields, where management zones address spatial variability. The SCMARL framework is hierarchical in nature, with a centralized coordinator agent at the top level and decentralized local agents at the second level. The coordinator agent makes daily binary irrigation decisions based on field-wide conditions, which are communicated to the local agents. Local agents determine appropriate irrigation amounts for specific management zones using local conditions. The framework employs state augmentation approach to handle non-stationarity in the local agents' environments. An extensive evaluation on a large-scale field in Lethbridge, Canada, compares the SCMARL approach with a learning-based multi-agent model predictive control scheduling approach, highlighting its enhanced performance, resulting in water conservation and improved Irrigation Water Use Efficiency (IWUE). Notably, the proposed approach achieved a 4.0% savings in irrigation water while enhancing the IWUE by 6.3%.
Title:
Generative Dataset Distillation Based on Diffusion Model
Authors: Duo Su, Junjie Hou, Guang Li, Ren Togo, Rui Song, Takahiro Ogawa, Miki Haseyama
Abstract
This paper presents our method for the generative track of The First Dataset Distillation Challenge at ECCV 2024. Since the diffusion model has become the mainstay of generative models because of its high-quality generative effects, we focus on distillation methods based on the diffusion model. Considering that the track can only generate a fixed number of images in 10 minutes using a generative model for CIFAR-100 and Tiny-ImageNet datasets, we need to use a generative model that can generate images at high speed. In this study, we proposed a novel generative dataset distillation method based on Stable Diffusion. Specifically, we use the SDXL-Turbo model which can generate images at high speed and quality. Compared to other diffusion models that can only generate images per class (IPC) = 1, our method can achieve an IPC = 10 for Tiny-ImageNet and an IPC = 20 for CIFAR-100, respectively. Additionally, to generate high-quality distilled datasets for CIFAR-100 and Tiny-ImageNet, we use the class information as text prompts and post data augmentation for the SDXL-Turbo model. Experimental results show the effectiveness of the proposed method, and we achieved third place in the generative track of the ECCV 2024 DD Challenge. Codes are available at this https URL.
Keyword: detection
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
intelligently coding for machine'' by cleverly leveraging the common sense of Large Multimodal Models (LMMs). We are motivated by the evidence that large language/multimodal models are powerful general-purpose semantics predictors for understanding the real world. Different from traditional image compression typically optimized for human eyes, the image coding for machines (ICM) framework we focus on requires the compressed bitstream to more comply with different downstream intelligent analysis tasks. To this end, we employ LMM to \textcolor{red}{tell codec what to compress}: 1) first utilize the powerful semantic understanding capability of LMMs w.r.t object grounding, identification, and importance ranking via prompts, to disentangle image content before compression, 2) and then based on these semantic priors we accordingly encode and transmit objects of the image in order with a structured bitstream. In this way, diverse vision benchmarks including image classification, object detection, instance segmentation, etc., can be well supported with such a semantically structured bitstream. We dub our method\textit{SDComp}'' for ``\textit{S}emantically \textit{D}isentangled \textit{Comp}ression'', and compare it with state-of-the-art codecs on a wide variety of different vision tasks. SDComp codec leads to more flexible reconstruction results, promised decoded visual quality, and a more generic/satisfactory intelligent task-supporting ability.Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Keyword: face recognition
There is no result
Keyword: augmentation
Title:
Title:
Title: