Abstract
With the rise of tiny IoT devices powered by machine learning (ML), many researchers have directed their focus toward compressing models to fit on tiny edge devices. Recent works have achieved remarkable success in compressing ML models for object detection and image classification on microcontrollers with small memory, e.g., 512kB SRAM. However, there remain many challenges prohibiting the deployment of ML systems that require high-resolution images. Due to fundamental limits in memory capacity for tiny IoT devices, it may be physically impossible to store large images without external hardware. To this end, we propose a high-resolution image scaling system for edge ML, called HiRISE, which is equipped with selective region-of-interest (ROI) capability leveraging analog in-sensor image scaling. Our methodology not only significantly reduces the peak memory requirements, but also achieves up to 17.7x reduction in data transfer and energy consumption.
Title:
PaveCap: The First Multimodal Framework for Comprehensive Pavement Condition Assessment with Dense Captioning and PCI Estimation
Abstract
This research introduces the first multimodal approach for pavement condition assessment, providing both quantitative Pavement Condition Index (PCI) predictions and qualitative descriptions. We introduce PaveCap, a novel framework for automated pavement condition assessment. The framework consists of two main parts: a Single-Shot PCI Estimation Network and a Dense Captioning Network. The PCI Estimation Network uses YOLOv8 for object detection, the Segment Anything Model (SAM) for zero-shot segmentation, and a four-layer convolutional neural network to predict PCI. The Dense Captioning Network uses a YOLOv8 backbone, a Transformer encoder-decoder architecture, and a convolutional feed-forward module to generate detailed descriptions of pavement conditions. To train and evaluate these networks, we developed a pavement dataset with bounding box annotations, textual annotations, and PCI values. The results of our PCI Estimation Network showed a strong positive correlation (0.70) between predicted and actual PCIs, demonstrating its effectiveness in automating condition assessment. Also, the Dense Captioning Network produced accurate pavement condition descriptions, evidenced by high BLEU (0.7445), GLEU (0.5893), and METEOR (0.7252) scores. Additionally, the dense captioning model handled complex scenarios well, even correcting some errors in the ground truth data. The framework developed here can greatly improve infrastructure management and decision18 making in pavement maintenance.
Title:
Multi-Scale and Detail-Enhanced Segment Anything Model for Salient Object Detection
Authors: Shixuan Gao, Pingping Zhang, Tianyu Yan, Huchuan Lu
Subjects: Subjects:
Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
Abstract
Salient Object Detection (SOD) aims to identify and segment the most prominent objects in images. Advanced SOD methods often utilize various Convolutional Neural Networks (CNN) or Transformers for deep feature extraction. However, these methods still deliver low performance and poor generalization in complex cases. Recently, Segment Anything Model (SAM) has been proposed as a visual fundamental model, which gives strong segmentation and generalization capabilities. Nonetheless, SAM requires accurate prompts of target objects, which are unavailable in SOD. Additionally, SAM lacks the utilization of multi-scale and multi-level information, as well as the incorporation of fine-grained details. To address these shortcomings, we propose a Multi-scale and Detail-enhanced SAM (MDSAM) for SOD. Specifically, we first introduce a Lightweight Multi-Scale Adapter (LMSA), which allows SAM to learn multi-scale information with very few trainable parameters. Then, we propose a Multi-Level Fusion Module (MLFM) to comprehensively utilize the multi-level information from the SAM's encoder. Finally, we propose a Detail Enhancement Module (DEM) to incorporate SAM with fine-grained details. Experimental results demonstrate the superior performance of our model on multiple SOD datasets and its strong generalization on other segmentation tasks. The source code is released at this https URL.
Title:
SAM2-Adapter: Evaluating & Adapting Segment Anything 2 in Downstream Tasks: Camouflage, Shadow, Medical Image Segmentation, and More
Abstract
The advent of large models, also known as foundation models, has significantly transformed the AI research landscape, with models like Segment Anything (SAM) achieving notable success in diverse image segmentation scenarios. Despite its advancements, SAM encountered limitations in handling some complex low-level segmentation tasks like camouflaged object and medical imaging. In response, in 2023, we introduced SAM-Adapter, which demonstrated improved performance on these challenging tasks. Now, with the release of Segment Anything 2 (SAM2), a successor with enhanced architecture and a larger training corpus, we reassess these challenges. This paper introduces SAM2-Adapter, the first adapter designed to overcome the persistent limitations observed in SAM2 and achieve new state-of-the-art (SOTA) results in specific downstream tasks including medical image segmentation, camouflaged (concealed) object detection, and shadow detection. SAM2-Adapter builds on the SAM-Adapter's strengths, offering enhanced generalizability and composability for diverse applications. We present extensive experimental results demonstrating SAM2-Adapter's effectiveness. We show the potential and encourage the research community to leverage the SAM2 model with our SAM2-Adapter for achieving superior segmentation outcomes. Code, pre-trained models, and data processing protocols are available at this http URL
Keyword: transformer
Title:
Image-to-LaTeX Converter for Mathematical Formulas and Text
Authors: Daniil Gurgurov, Aleksey Morshnev
Subjects: Subjects:
Computation and Language (cs.CL); Computer Vision and Pattern Recognition (cs.CV)
Abstract
In this project, we train a vision encoder-decoder model to generate LaTeX code from images of mathematical formulas and text. Utilizing a diverse collection of image-to-LaTeX data, we build two models: a base model with a Swin Transformer encoder and a GPT-2 decoder, trained on machine-generated images, and a fine-tuned version enhanced with Low-Rank Adaptation (LoRA) trained on handwritten formulas. We then compare the BLEU performance of our specialized model on a handwritten test set with other similar models, such as Pix2Text, TexTeller, and Sumen. Through this project, we contribute open-source models for converting images to LaTeX and provide from-scratch code for building these models with distributed training and GPU optimizations.
Title:
Synthesizing Abstract Transformers for Reduced-Product Domains
Authors: Pankaj Kumar Kalita, Thomas Reps, Subhajit Roy
Abstract
Recently, we showed how to apply program-synthesis techniques to create abstract transformers in a user-provided domain-specific language (DSL) L (i.e., ''L-transformers"). However, we found that the algorithm of Kalita et al. does not succeed when applied to reduced-product domains: the need to synthesize transformers for all of the domains simultaneously blows up the search space. Because reduced-product domains are an important device for improving the precision of abstract interpretation, in this paper, we propose an algorithm to synthesize reduced L-transformers $\langle f_1^{\sharp R}, f_2^{\sharp R},..., f_n^{\sharp R}\rangle$ for a product domain $A_1 \times A_2 \times \ldots \times A_n$ , using multiple DSLs: $\mathcal{L} = \langle \mathcal{L}_1 , \mathcal{L}_2, ... , \mathcal{L}_n \rangle$. Synthesis of reduced-product transformers is quite challenging: first, the synthesis task has to tackle an increased ''feature set" because each component transformer now has access to the abstract inputs from all component domains in the product. Second, to ensure that the product transformer is maximally precise, the synthesis task needs to arrange for the component transformers to cooperate with each other. We implemented our algorithm in a tool, Amurth2, and used it to synthesize abstract transformers for two product domains -- SAFE and JSAI -- available within the SAFEstr framework for JavaScript program analysis. For four of the six operations supported by SAFEstr, Amurth2 synthesizes more precise abstract transformers than the manually written ones available in SAFEstr.
Title:
Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters
Authors: Vasudev Shyam, Jonathan Pilault, Emily Shepperd, Quentin Anthony, Beren Millidge
Subjects: Subjects:
Machine Learning (cs.LG); Computation and Language (cs.CL)
Abstract
Self-attention is the core mathematical operation of modern transformer architectures and is also a significant computational bottleneck due to its quadratic complexity in the sequence length. In this work, we derive the scalar energy function whose gradient computes the self-attention block, thus elucidating the theoretical underpinnings of self-attention, providing a Bayesian interpretation of the operation and linking it closely with energy-based models such as Hopfield Networks. Moreover, due to this formulation, we discover that we can use efficient and optimized automatic-differentiation techniques to derive a highly efficient Tree Attention algorithm to compute the gradient of the energy and hence self-attention. Our formulation reveals that the reduction across the sequence axis can be efficiently computed in parallel through a tree reduction. Our algorithm, for parallelizing attention computation across multiple GPUs, enables cross-device decoding to be performed asymptotically faster (up to 8x faster) than alternative approaches such as Ring Attention, while also requiring significantly less communication volume and incurring 2x less peak memory. Our code is publicly available here: \url{this https URL}
Title:
PaveCap: The First Multimodal Framework for Comprehensive Pavement Condition Assessment with Dense Captioning and PCI Estimation
Abstract
This research introduces the first multimodal approach for pavement condition assessment, providing both quantitative Pavement Condition Index (PCI) predictions and qualitative descriptions. We introduce PaveCap, a novel framework for automated pavement condition assessment. The framework consists of two main parts: a Single-Shot PCI Estimation Network and a Dense Captioning Network. The PCI Estimation Network uses YOLOv8 for object detection, the Segment Anything Model (SAM) for zero-shot segmentation, and a four-layer convolutional neural network to predict PCI. The Dense Captioning Network uses a YOLOv8 backbone, a Transformer encoder-decoder architecture, and a convolutional feed-forward module to generate detailed descriptions of pavement conditions. To train and evaluate these networks, we developed a pavement dataset with bounding box annotations, textual annotations, and PCI values. The results of our PCI Estimation Network showed a strong positive correlation (0.70) between predicted and actual PCIs, demonstrating its effectiveness in automating condition assessment. Also, the Dense Captioning Network produced accurate pavement condition descriptions, evidenced by high BLEU (0.7445), GLEU (0.5893), and METEOR (0.7252) scores. Additionally, the dense captioning model handled complex scenarios well, even correcting some errors in the ground truth data. The framework developed here can greatly improve infrastructure management and decision18 making in pavement maintenance.
Title:
M2EF-NNs: Multimodal Multi-instance Evidence Fusion Neural Networks for Cancer Survival Prediction
Abstract
Accurate cancer survival prediction is crucial for assisting clinical doctors in formulating treatment plans. Multimodal data, including histopathological images and genomic data, offer complementary and comprehensive information that can greatly enhance the accuracy of this task. However, the current methods, despite yielding promising results, suffer from two notable limitations: they do not effectively utilize global context and disregard modal uncertainty. In this study, we put forward a neural network model called M2EF-NNs, which leverages multimodal and multi-instance evidence fusion techniques for accurate cancer survival prediction. Specifically, to capture global information in the images, we use a pre-trained Vision Transformer (ViT) model to obtain patch feature embeddings of histopathological images. Then, we introduce a multimodal attention module that uses genomic embeddings as queries and learns the co-attention mapping between genomic and histopathological images to achieve an early interaction fusion of multimodal information and better capture their correlations. Subsequently, we are the first to apply the Dempster-Shafer evidence theory (DST) to cancer survival prediction. We parameterize the distribution of class probabilities using the processed multimodal features and introduce subjective logic to estimate the uncertainty associated with different modalities. By combining with the Dempster-Shafer theory, we can dynamically adjust the weights of class probabilities after multimodal fusion to achieve trusted survival prediction. Finally, Experimental validation on the TCGA datasets confirms the significant improvements achieved by our proposed method in cancer survival prediction and enhances the reliability of the model.
Title:
MultiColor: Image Colorization by Learning from Multiple Color Spaces
Authors: Xiangcheng Du, Zhao Zhou, Yanlong Wang, Zhuoyao Wang, Yingbin Zheng, Cheng Jin
Subjects: Subjects:
Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
Abstract
Deep networks have shown impressive performance in the image restoration tasks, such as image colorization. However, we find that previous approaches rely on the digital representation from single color model with a specific mapping function, a.k.a., color space, during the colorization pipeline. In this paper, we first investigate the modeling of different color spaces, and find each of them exhibiting distinctive characteristics with unique distribution of colors. The complementarity among multiple color spaces leads to benefits for the image colorization task. We present MultiColor, a new learning-based approach to automatically colorize grayscale images that combines clues from multiple color spaces. Specifically, we employ a set of dedicated colorization modules for individual color space. Within each module, a transformer decoder is first employed to refine color query embeddings and then a color mapper produces color channel prediction using the embeddings and semantic features. With these predicted color channels representing various color spaces, a complementary network is designed to exploit the complementarity and generate pleasing and reasonable colorized images. We conduct extensive experiments on real-world datasets, and the results demonstrate superior performance over the state-of-the-arts.
Title:
F1tenth Autonomous Racing With Offline Reinforcement Learning Methods
Authors: Prajwal Koirala, Cody Fleming
Subjects: Subjects:
Robotics (cs.RO); Systems and Control (eess.SY)
Abstract
Autonomous racing serves as a critical platform for evaluating automated driving systems and enhancing vehicle mobility intelligence. This work investigates offline reinforcement learning methods to train agents within the dynamic F1tenth racing environment. The study begins by exploring the challenges of online training in the Austria race track environment, where agents consistently fail to complete the laps. Consequently, this research pivots towards an offline strategy, leveraging `expert' demonstration dataset to facilitate agent training. A waypoint-based suboptimal controller is developed to gather data with successful lap episodes. This data is then employed to train offline learning-based algorithms, with a subsequent analysis of the agents' cross-track performance, evaluating their zero-shot transferability from seen to unseen scenarios and their capacity to adapt to changes in environment dynamics. Beyond mere algorithm benchmarking in autonomous racing scenarios, this study also introduces and describes the machinery of our return-conditioned decision tree-based policy, comparing its performance with methods that employ fully connected neural networks, Transformers, and Diffusion Policies and highlighting some insights into method selection for training autonomous agents in driving interactions.
Title:
Attention Mechanism and Context Modeling System for Text Mining Machine Translation
Authors: Shi Bo, Yuwei Zhang, Junming Huang, Sitong Liu, Zexi Chen, Zizheng Li
Subjects: Subjects:
Computation and Language (cs.CL); Artificial Intelligence (cs.AI)
Abstract
This paper advances a novel architectural schema anchored upon the Transformer paradigm and innovatively amalgamates the K-means categorization algorithm to augment the contextual apprehension capabilities of the schema. The transformer model performs well in machine translation tasks due to its parallel computing power and multi-head attention mechanism. However, it may encounter contextual ambiguity or ignore local features when dealing with highly complex language structures. To circumvent this constraint, this exposition incorporates the K-Means algorithm, which is used to stratify the lexis and idioms of the input textual matter, thereby facilitating superior identification and preservation of the local structure and contextual intelligence of the language. The advantage of this combination is that K-Means can automatically discover the topic or concept regions in the text, which may be directly related to translation quality. Consequently, the schema contrived herein enlists K-Means as a preparatory phase antecedent to the Transformer and recalibrates the multi-head attention weights to assist in the discrimination of lexis and idioms bearing analogous semantics or functionalities. This ensures the schema accords heightened regard to the contextual intelligence embodied by these clusters during the training phase, rather than merely focusing on locational intelligence.
Title:
LLDif: Diffusion Models for Low-light Emotion Recognition
Abstract
This paper introduces LLDif, a novel diffusion-based facial expression recognition (FER) framework tailored for extremely low-light (LL) environments. Images captured under such conditions often suffer from low brightness and significantly reduced contrast, presenting challenges to conventional methods. These challenges include poor image quality that can significantly reduce the accuracy of emotion recognition. LLDif addresses these issues with a novel two-stage training process that combines a Label-aware CLIP (LA-CLIP), an embedding prior network (PNET), and a transformer-based network adept at handling the noise of low-light images. The first stage involves LA-CLIP generating a joint embedding prior distribution (EPD) to guide the LLformer in label recovery. In the second stage, the diffusion model (DM) refines the EPD inference, ultilising the compactness of EPD for precise predictions. Experimental evaluations on various LL-FER datasets have shown that LLDif achieves competitive performance, underscoring its potential to enhance FER applications in challenging lighting conditions.
Title:
Cluster-Wide Task Slowdown Detection in Cloud System
Abstract
Slow task detection is a critical problem in cloud operation and maintenance since it is highly related to user experience and can bring substantial liquidated damages. Most anomaly detection methods detect it from a single-task aspect. However, considering millions of concurrent tasks in large-scale cloud computing clusters, it becomes impractical and inefficient. Moreover, single-task slowdowns are very common and do not necessarily indicate a malfunction of a cluster due to its violent fluctuation nature in a virtual environment. Thus, we shift our attention to cluster-wide task slowdowns by utilizing the duration time distribution of tasks across a cluster, so that the computation complexity is not relevant to the number of tasks. The task duration time distribution often exhibits compound periodicity and local exceptional fluctuations over time. Though transformer-based methods are one of the most powerful methods to capture these time series normal variation patterns, we empirically find and theoretically explain the flaw of the standard attention mechanism in reconstructing subperiods with low amplitude when dealing with compound periodicity. To tackle these challenges, we propose SORN (i.e., Skimming Off subperiods in descending amplitude order and Reconstructing Non-slowing fluctuation), which consists of a Skimming Attention mechanism to reconstruct the compound periodicity and a Neural Optimal Transport module to distinguish cluster-wide slowdowns from other exceptional fluctuations. Furthermore, since anomalies in the training set are inevitable in a practical scenario, we propose a picky loss function, which adaptively assigns higher weights to reliable time slots in the training set. Extensive experiments demonstrate that SORN outperforms state-of-the-art methods on multiple real-world industrial datasets.
Title:
Scalable Transformer for High Dimensional Multivariate Time Series Forecasting
Abstract
Deep models for Multivariate Time Series (MTS) forecasting have recently demonstrated significant success. Channel-dependent models capture complex dependencies that channel-independent models cannot capture. However, the number of channels in real-world applications outpaces the capabilities of existing channel-dependent models, and contrary to common expectations, some models underperform the channel-independent models in handling high-dimensional data, which raises questions about the performance of channel-dependent models. To address this, our study first investigates the reasons behind the suboptimal performance of these channel-dependent models on high-dimensional MTS data. Our analysis reveals that two primary issues lie in the introduced noise from unrelated series that increases the difficulty of capturing the crucial inter-channel dependencies, and challenges in training strategies due to high-dimensional data. To address these issues, we propose STHD, the Scalable Transformer for High-Dimensional Multivariate Time Series Forecasting. STHD has three components: a) Relation Matrix Sparsity that limits the noise introduced and alleviates the memory issue; b) ReIndex applied as a training strategy to enable a more flexible batch size setting and increase the diversity of training data; and c) Transformer that handles 2-D inputs and captures channel dependencies. These components jointly enable STHD to manage the high-dimensional MTS while maintaining computational feasibility. Furthermore, experimental results show STHD's considerable improvement on three high-dimensional datasets: Crime-Chicago, Wiki-People, and Traffic. The source code and dataset are publicly available this https URL.
Title:
UHNet: An Ultra-Lightweight and High-Speed Edge Detection Network
Authors: Fuzhang Li, Chuan Lin
Subjects: Subjects:
Computer Vision and Pattern Recognition (cs.CV)
Abstract
Edge detection is crucial in medical image processing, enabling precise extraction of structural information to support lesion identification and image analysis. Traditional edge detection models typically rely on complex Convolutional Neural Networks and Vision Transformer architectures. Due to their numerous parameters and high computational demands, these models are limited in their application on resource-constrained devices. This paper presents an ultra-lightweight edge detection model (UHNet), characterized by its minimal parameter count, rapid computation speed, negligible of pre-training costs, and commendable performance. UHNet boasts impressive performance metrics with 42.3k parameters, 166 FPS, and 0.79G FLOPs. By employing an innovative feature extraction module and optimized residual connection method, UHNet significantly reduces model complexity and computational requirements. Additionally, a lightweight feature fusion strategy is explored, enhancing detection accuracy. Experimental results on the BSDS500, NYUD, and BIPED datasets validate that UHNet achieves remarkable edge detection performance while maintaining high efficiency. This work not only provides new insights into the design of lightweight edge detection models but also demonstrates the potential and application prospects of the UHNet model in engineering applications such as medical image processing. The codes are available at this https URL
Title:
LaDiMo: Layer-wise Distillation Inspired MoEfier
Authors: Sungyoon Kim, Youngjun Kim, Kihyo Moon, Minsung Jang
Subjects: Subjects:
Computation and Language (cs.CL)
Abstract
The advent of large language models has revolutionized natural language processing, but their increasing complexity has led to substantial training costs, resource demands, and environmental impacts. In response, sparse Mixture-of-Experts (MoE) models have emerged as a promising alternative to dense models. Since training MoE models from scratch can be prohibitively expensive, recent studies have explored leveraging knowledge from pre-trained non-MoE models. However, existing approaches have limitations, such as requiring significant hardware resources and data. We propose a novel algorithm, LaDiMo, which efficiently converts a Transformer-based non-MoE model into a MoE model with minimal additional training cost. LaDiMo consists of two stages: layer-wise expert construction and routing policy decision. By harnessing the concept of Knowledge Distillation, we compress the model and rapidly recover its performance. Furthermore, we develop an adaptive router that optimizes inference efficiency by profiling the distribution of routing weights and determining a layer-wise policy that balances accuracy and latency. We demonstrate the effectiveness of our method by converting the LLaMA2-7B model to a MoE model using only 100K tokens, reducing activated parameters by over 20% while keeping accuracy. Our approach offers a flexible and efficient solution for building and deploying MoE models.
Title:
TheGlueNote: Learned Representations for Robust and Flexible Note Alignment
Abstract
Note alignment refers to the task of matching individual notes of two versions of the same symbolically encoded piece. Methods addressing this task commonly rely on sequence alignment algorithms such as Hidden Markov Models or Dynamic Time Warping (DTW) applied directly to note or onset sequences. While successful in many cases, such methods struggle with large mismatches between the versions. In this work, we learn note-wise representations from data augmented with various complex mismatch cases, e.g. repeats, skips, block insertions, and long trills. At the heart of our approach lies a transformer encoder network - TheGlueNote - which predicts pairwise note similarities for two 512 note subsequences. We postprocess the predicted similarities using flavors of weightedDTW and pitch-separated onsetDTW to retrieve note matches for two sequences of arbitrary length. Our approach performs on par with the state of the art in terms of note alignment accuracy, is considerably more robust to version mismatches, and works directly on any pair of MIDI files.
Title:
Multi-Scale and Detail-Enhanced Segment Anything Model for Salient Object Detection
Authors: Shixuan Gao, Pingping Zhang, Tianyu Yan, Huchuan Lu
Subjects: Subjects:
Computer Vision and Pattern Recognition (cs.CV); Multimedia (cs.MM)
Abstract
Salient Object Detection (SOD) aims to identify and segment the most prominent objects in images. Advanced SOD methods often utilize various Convolutional Neural Networks (CNN) or Transformers for deep feature extraction. However, these methods still deliver low performance and poor generalization in complex cases. Recently, Segment Anything Model (SAM) has been proposed as a visual fundamental model, which gives strong segmentation and generalization capabilities. Nonetheless, SAM requires accurate prompts of target objects, which are unavailable in SOD. Additionally, SAM lacks the utilization of multi-scale and multi-level information, as well as the incorporation of fine-grained details. To address these shortcomings, we propose a Multi-scale and Detail-enhanced SAM (MDSAM) for SOD. Specifically, we first introduce a Lightweight Multi-Scale Adapter (LMSA), which allows SAM to learn multi-scale information with very few trainable parameters. Then, we propose a Multi-Level Fusion Module (MLFM) to comprehensively utilize the multi-level information from the SAM's encoder. Finally, we propose a Detail Enhancement Module (DEM) to incorporate SAM with fine-grained details. Experimental results demonstrate the superior performance of our model on multiple SOD datasets and its strong generalization on other segmentation tasks. The source code is released at this https URL.
Title:
Towards Explainable Network Intrusion Detection using Large Language Models
Authors: Paul R. B. Houssel, Priyanka Singh, Siamak Layeghy, Marius Portmann
Subjects: Subjects:
Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Networking and Internet Architecture (cs.NI)
Abstract
Large Language Models (LLMs) have revolutionised natural language processing tasks, particularly as chat agents. However, their applicability to threat detection problems remains unclear. This paper examines the feasibility of employing LLMs as a Network Intrusion Detection System (NIDS), despite their high computational requirements, primarily for the sake of explainability. Furthermore, considerable resources have been invested in developing LLMs, and they may offer utility for NIDS. Current state-of-the-art NIDS rely on artificial benchmarking datasets, resulting in skewed performance when applied to real-world networking environments. Therefore, we compare the GPT-4 and LLama3 models against traditional architectures and transformer-based models to assess their ability to detect malicious NetFlows without depending on artificially skewed datasets, but solely on their vast pre-trained acquired knowledge. Our results reveal that, although LLMs struggle with precise attack detection, they hold significant potential for a path towards explainable NIDS. Our preliminary exploration shows that LLMs are unfit for the detection of Malicious NetFlows. Most promisingly, however, these exhibit significant potential as complementary agents in NIDS, particularly in providing explanations and aiding in threat response when integrated with Retrieval Augmented Generation (RAG) and function calling capabilities.
Title:
Deeploy: Enabling Energy-Efficient Deployment of Small Language Models On Heterogeneous Microcontrollers
Authors: Moritz Scherer, Luka Macan, Victor Jung, Philip Wiese, Luca Bompani, Alessio Burrello, Francesco Conti, Luca Benini
Abstract
With the rise of Embodied Foundation Models (EFMs), most notably Small Language Models (SLMs), adapting Transformers for edge applications has become a very active field of research. However, achieving end-to-end deployment of SLMs on microcontroller (MCU)-class chips without high-bandwidth off-chip main memory access is still an open challenge. In this paper, we demonstrate high-efficiency end-to-end SLM deployment on a multicore RISC-V (RV32) MCU augmented with ML instruction extensions and a hardware neural processing unit (NPU). To automate the exploration of the constrained, multi-dimensional memory vs. computation tradeoffs involved in aggressive SLM deployment on heterogeneous (multicore+NPU) resources, we introduce Deeploy, a novel Deep Neural Network (DNN) compiler, which generates highly-optimized C code requiring minimal runtime support. We demonstrate that Deeploy generates end-to-end code for executing SLMs, fully exploiting the RV32 cores' instruction extensions and the NPU: We achieve leading-edge energy and throughput of \SI{490}{\micro\joule \per Token}, at \SI{340}{Token \per \second} for an SLM trained on the TinyStories dataset, running for the first time on an MCU-class device without external memory.
Title:
Crowd Intelligence for Early Misinformation Prediction on Social Media
Abstract
Misinformation spreads rapidly on social media, causing serious damage by influencing public opinion, promoting dangerous behavior, or eroding trust in reliable sources. It spreads too fast for traditional fact-checking, stressing the need for predictive methods. We introduce CROWDSHIELD, a crowd intelligence-based method for early misinformation prediction. We hypothesize that the crowd's reactions to misinformation reveal its accuracy. Furthermore, we hinge upon exaggerated assertions/claims and replies with particular positions/stances on the source post within a conversation thread. We employ Q-learning to capture the two dimensions -- stances and claims. We utilize deep Q-learning due to its proficiency in navigating complex decision spaces and effectively learning network properties. Additionally, we use a transformer-based encoder to develop a comprehensive understanding of both content and context. This multifaceted approach helps ensure the model pays attention to user interaction and stays anchored in the communication's content. We propose MIST, a manually annotated misinformation detection Twitter corpus comprising nearly 200 conversation threads with more than 14K replies. In experiments, CROWDSHIELD outperformed ten baseline systems, achieving an improvement of ~4% macro-F1 score. We conduct an ablation study and error analysis to validate our proposed model's performance. The source code and dataset are available at this https URL.
Title:
How Transformers Utilize Multi-Head Attention in In-Context Learning? A Case Study on Sparse Linear Regression
Abstract
Despite the remarkable success of transformer-based models in various real-world tasks, their underlying mechanisms remain poorly understood. Recent studies have suggested that transformers can implement gradient descent as an in-context learner for linear regression problems and have developed various theoretical analyses accordingly. However, these works mostly focus on the expressive power of transformers by designing specific parameter constructions, lacking a comprehensive understanding of their inherent working mechanisms post-training. In this study, we consider a sparse linear regression problem and investigate how a trained multi-head transformer performs in-context learning. We experimentally discover that the utilization of multi-heads exhibits different patterns across layers: multiple heads are utilized and essential in the first layer, while usually only a single head is sufficient for subsequent layers. We provide a theoretical explanation for this observation: the first layer preprocesses the context data, and the following layers execute simple optimization steps based on the preprocessed context. Moreover, we demonstrate that such a preprocess-then-optimize algorithm can significantly outperform naive gradient descent and ridge regression algorithms. Further experimental results support our explanations. Our findings offer insights into the benefits of multi-head attention and contribute to understanding the more intricate mechanisms hidden within trained transformers.
Title:
Towards Resilient and Efficient LLMs: A Comparative Study of Efficiency, Performance, and Adversarial Robustness
Authors: Xiaojing Fan, Chunliang Tao
Subjects: Subjects:
Computation and Language (cs.CL)
Abstract
With the increasing demand for practical applications of Large Language Models (LLMs), many attention-efficient models have been developed to balance performance and computational cost. However, the adversarial robustness of these models remains under-explored. In this work, we design a framework to investigate the trade-off between efficiency, performance, and adversarial robustness of LLMs by comparing three prominent models with varying levels of complexity and efficiency -- Transformer++, Gated Linear Attention (GLA) Transformer, and MatMul-Free LM -- utilizing the GLUE and AdvGLUE datasets. The AdvGLUE dataset extends the GLUE dataset with adversarial samples designed to challenge model robustness. Our results show that while the GLA Transformer and MatMul-Free LM achieve slightly lower accuracy on GLUE tasks, they demonstrate higher efficiency and either superior or comparative robustness on AdvGLUE tasks compared to Transformer++ across different attack levels. These findings highlight the potential of simplified architectures to achieve a compelling balance between efficiency, performance, and adversarial robustness, offering valuable insights for applications where resource constraints and resilience to adversarial attacks are critical.
Title:
Towards High-resolution 3D Anomaly Detection via Group-Level Feature Contrastive Learning
Authors: Hongze Zhu, Guoyang Xie, Chengbin Hou, Tao Dai, Can Gao, Jinbao Wang, Linlin Shen
Subjects: Subjects:
Computer Vision and Pattern Recognition (cs.CV)
Abstract
High-resolution point clouds~(HRPCD) anomaly detection~(AD) plays a critical role in precision machining and high-end equipment manufacturing. Despite considerable 3D-AD methods that have been proposed recently, they still cannot meet the requirements of the HRPCD-AD task. There are several challenges: i) It is difficult to directly capture HRPCD information due to large amounts of points at the sample level; ii) The advanced transformer-based methods usually obtain anisotropic features, leading to degradation of the representation; iii) The proportion of abnormal areas is very small, which makes it difficult to characterize. To address these challenges, we propose a novel group-level feature-based network, called Group3AD, which has a significantly efficient representation ability. First, we design an Intercluster Uniformity Network~(IUN) to present the mapping of different groups in the feature space as several clusters, and obtain a more uniform distribution between clusters representing different parts of the point clouds in the feature space. Then, an Intracluster Alignment Network~(IAN) is designed to encourage groups within the cluster to be distributed tightly in the feature space. In addition, we propose an Adaptive Group-Center Selection~(AGCS) based on geometric information to improve the pixel density of potential anomalous regions during inference. The experimental results verify the effectiveness of our proposed Group3AD, which surpasses Reg3D-AD by the margin of 5\% in terms of object-level AUROC on Real3D-AD. We provide the code and supplementary information on our website: this https URL.
Title:
Transformer Explainer: Interactive Learning of Text-Generative Models
Authors: Aeree Cho, Grace C. Kim, Alexander Karpekov, Alec Helbling, Zijie J. Wang, Seongmin Lee, Benjamin Hoover, Duen Horng Chau
Subjects: Subjects:
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Human-Computer Interaction (cs.HC)
Abstract
Transformers have revolutionized machine learning, yet their inner workings remain opaque to many. We present Transformer Explainer, an interactive visualization tool designed for non-experts to learn about Transformers through the GPT-2 model. Our tool helps users understand complex Transformer concepts by integrating a model overview and enabling smooth transitions across abstraction levels of mathematical operations and model structures. It runs a live GPT-2 instance locally in the user's browser, empowering users to experiment with their own input and observe in real-time how the internal components and parameters of the Transformer work together to predict the next tokens. Our tool requires no installation or special hardware, broadening the public's education access to modern generative AI techniques. Our open-sourced tool is available at this https URL. A video demo is available at this https URL.
Keyword: scene understanding
There is no result
Keyword: visual reasoning
Title:
ArtVLM: Attribute Recognition Through Vision-Based Prefix Language Modeling
Authors: William Y. Zhu, Keren Ye, Junjie Ke, Jiahui Yu, Leonidas Guibas, Peyman Milanfar, Feng Yang
Abstract
Recognizing and disentangling visual attributes from objects is a foundation to many computer vision applications. While large vision language representations like CLIP had largely resolved the task of zero-shot object recognition, zero-shot visual attribute recognition remains a challenge because CLIP's contrastively-learned vision-language representation cannot effectively capture object-attribute dependencies. In this paper, we target this weakness and propose a sentence generation-based retrieval formulation for attribute recognition that is novel in 1) explicitly modeling a to-be-measured and retrieved object-attribute relation as a conditional probability graph, which converts the recognition problem into a dependency-sensitive language-modeling problem, and 2) applying a large pretrained Vision-Language Model (VLM) on this reformulation and naturally distilling its knowledge of image-object-attribute relations to use towards attribute recognition. Specifically, for each attribute to be recognized on an image, we measure the visual-conditioned probability of generating a short sentence encoding the attribute's relation to objects on the image. Unlike contrastive retrieval, which measures likelihood by globally aligning elements of the sentence to the image, generative retrieval is sensitive to the order and dependency of objects and attributes in the sentence. We demonstrate through experiments that generative retrieval consistently outperforms contrastive retrieval on two visual reasoning datasets, Visual Attribute in the Wild (VAW), and our newly-proposed Visual Genome Attribute Ranking (VGARank).
Keyword: human object interaction
There is no result
Keyword: visual relation detection
There is no result
Keyword: object detection
Title:
Title:
Title:
Title:
Keyword: transformer
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Title:
Keyword: scene understanding
There is no result
Keyword: visual reasoning
Title: