PyTorchKR Papers

최신 논문

총 9,607편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2410.23230

Aligning Audio-Visual Joint Representations with an Agentic Workflow

Visual content and accompanied audio signals naturally formulate a joint representation to improve audio-visual (AV) related applications. While studies develop various AV representation learning frameworks, the importance of AV data alignment is usually undermined for achieving high-quality representation. We observe that an audio signal may contain background noise interference. Also, non-synchronization may appear between audio and video streams. These non-strict data alignment limits representation quality and downgrade application performance. In this paper, we propose to improve AV joint representations from a data-centric perspective by aligning audio signals to visual data. Our alignment is conducted in an agentic workflow controlled by an LLM-based assistant named AVAgent. For each input AV data pair, our AVAgent uses a multi-modal LLM to convert audio and visual data into language descriptions separately (i.e., tool use). Then, AVAgent reasons whether this paired data is aligned well and plans to edit the audio signal if needed (i.e., planning). The audio editing is executed by predefined actions that filter noise or augment data. Moreover, we use a VLM to evaluate how modified audio signals match the visual content and provide feedback to AVAgent (i.e., reflection). The tool use, planning, and reflection steps operate cyclically to become an agentic workflow where audio signals are gradually aligned to visual content. To this end, existing methods can directly leverage the aligned AV data via our agentic workflow to improve AV joint representations. The experimental results comprehensively demonstrate the state-of-the-art performance of the proposed approach against previous baselines in diverse downstream tasks.

arXiv:2410.23126

Provably Optimal Memory Capacity for Modern Hopfield Models: Transformer-Compatible Dense Associative Memories as Spherical Codes

We study the optimal memorization capacity of modern Hopfield models and Kernelized Hopfield Models (KHMs), a transformer-compatible class of Dense Associative Memories. We present a tight analysis by establishing a connection between the memory configuration of KHMs and spherical codes from information theory. Specifically, we treat the stored memory set as a specialized spherical code. This enables us to cast the memorization problem in KHMs into a point arrangement problem on a hypersphere. We show that the optimal capacity of KHMs occurs when the feature space allows memories to form an optimal spherical code. This unique perspective leads to: (i) An analysis of how KHMs achieve optimal memory capacity, and identify corresponding necessary conditions. Importantly, we establish an upper capacity bound that matches the well-known exponential lower bound in the literature. This provides the first tight and optimal asymptotic memory capacity for modern Hopfield models. (ii) A sub-linear time algorithm $\mathtt{U}\text{-}\mathtt{Hop}$+ to reach KHMs' optimal capacity. (iii) An analysis of the scaling behavior of the required feature dimension relative to the number of stored memories. These efforts improve both the retrieval capability of KHMs and the representation learning of corresponding transformers. Experimentally, we provide thorough numerical results to back up theoretical findings.

arXiv:2410.21803

SimSiam Naming Game: A Unified Approach for Representation Learning and Emergent Communication

Emergent communication, driven by generative models, enables agents to develop a shared language for describing their individual views of the same objects through interactions. Meanwhile, self-supervised learning (SSL), particularly SimSiam, uses discriminative representation learning to make representations of augmented views of the same data point closer in the representation space. Building on the prior work of VI-SimSiam, which incorporates a generative and Bayesian perspective into the SimSiam framework via variational inference (VI) interpretation, we propose SimSiam+VAE, a unified approach for both representation learning and emergent communication. SimSiam+VAE integrates a variational autoencoder (VAE) into the predictor of the SimSiam network to enhance representation learning and capture uncertainty. Experimental results show that SimSiam+VAE outperforms both SimSiam and VI-SimSiam. We further extend this model into a communication framework called the SimSiam Naming Game (SSNG), which applies the generative and Bayesian approach based on VI to develop internal representations and emergent language, while utilizing the discriminative process of SimSiam to facilitate mutual understanding between agents. In experiments with established models, despite the dynamic alternation of agent roles during interactions, SSNG demonstrates comparable performance to the referential game and slightly outperforms the Metropolis-Hastings naming game.

arXiv:2410.22026

Enhance Hyperbolic Representation Learning via Second-order Pooling

Hyperbolic representation learning is well known for its ability to capture hierarchical information. However, the distance between samples from different levels of hierarchical classes can be required large. We reveal that the hyperbolic discriminant objective forces the backbone to capture this hierarchical information, which may inevitably increase the Lipschitz constant of the backbone. This can hinder the full utilization of the backbone's generalization ability. To address this issue, we introduce second-order pooling into hyperbolic representation learning, as it naturally increases the distance between samples without compromising the generalization ability of the input features. In this way, the Lipschitz constant of the backbone does not necessarily need to be large. However, current off-the-shelf low-dimensional bilinear pooling methods cannot be directly employed in hyperbolic representation learning because they inevitably reduce the distance expansion capability. To solve this problem, we propose a kernel approximation regularization, which enables the low-dimensional bilinear features to approximate the kernel function well in low-dimensional space. Finally, we conduct extensive experiments on graph-structured datasets to demonstrate the effectiveness of the proposed method.

arXiv:2410.22325

Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets

The pre-training of visual representations has enhanced the efficiency of robot learning. Due to the lack of large-scale in-domain robotic datasets, prior works utilize in-the-wild human videos to pre-train robotic visual representation. Despite their promising results, representations from human videos are inevitably subject to distribution shifts and lack the dynamics information crucial for task completion. We first evaluate various pre-trained representations in terms of their correlation to the downstream robotic manipulation tasks (i.e., manipulation centricity). Interestingly, we find that the "manipulation centricity" is a strong indicator of success rates when applied to downstream tasks. Drawing from these findings, we propose Manipulation Centric Representation (MCR), a foundation representation learning framework capturing both visual features and the dynamics information such as actions and proprioceptions of manipulation tasks to improve manipulation centricity. Specifically, we pre-train a visual encoder on the DROID robotic dataset and leverage motion-relevant data such as robot proprioceptive states and actions. We introduce a novel contrastive loss that aligns visual observations with the robot's proprioceptive state-action dynamics, combined with a behavior cloning (BC)-like actor loss to predict actions during pre-training, along with a time contrastive loss. Empirical results across 4 simulation domains with 20 tasks verify that MCR outperforms the strongest baseline method by 14.8%. Moreover, MCR boosts the performance of data-efficient learning with a UR5e arm on 3 real-world tasks by 76.9%. Project website: https://robots-pretrain-robots.github.io/.

arXiv:2410.20542공식 코드 · Papers with Code

PaPaGei: Open Foundation Models for Optical Physiological Signals

Photoplethysmography (PPG) is the leading non-invasive technique for monitoring biosignals and cardiovascular health, with widespread adoption in both clinical settings and consumer wearable devices. While machine learning models trained on PPG signals have shown promise, they tend to be task-specific and struggle with generalization. Current research is limited by the use of single-device datasets, insufficient exploration of out-of-domain generalization, and a lack of publicly available models, which hampers reproducibility. To address these limitations, we present PaPaGei, the first open foundation model for PPG signals. The model is pre-trained on over 57,000 hours of data, comprising 20 million unlabeled PPG segments from publicly available datasets. We introduce a novel representation learning approach that leverages domain knowledge of PPG signal morphology across individuals, enabling the capture of richer representations compared to traditional contrastive learning methods. We evaluate PaPaGei against state-of-the-art time-series foundation models and self-supervised learning benchmarks across 20 tasks from 10 diverse datasets, spanning cardiovascular health, sleep disorders, pregnancy monitoring, and wellbeing assessment. Our model demonstrates superior performance, improving classification and regression metrics by 6.3% and 2.9% respectively in at least 14 tasks. Notably, PaPaGei achieves these results while being more data- and parameter-efficient, outperforming models that are 70x larger. Beyond accuracy, we examine model robustness across different skin tones, establishing a benchmark for bias evaluation in future models. PaPaGei can serve as both a feature extractor and an encoder for multimodal models, opening up new opportunities for multimodal health monitoring.

arXiv:2410.19877

Foundation Models in Electrocardiogram: A Review

The electrocardiogram (ECG) is ubiquitous across various healthcare domains, such as cardiac arrhythmia detection and sleep monitoring, making ECG analysis critically essential. Traditional deep learning models for ECG are task-specific, with a narrow scope of functionality and limited generalization capabilities. Recently, foundation models (FMs), also known as large pre-training models, have fundamentally reshaped the scheme of model design and representation learning, enhancing the performance across a variety of downstream tasks. This success has drawn interest in the exploration of FMs to address ECG-based medical challenges concurrently. This survey provides a timely, comprehensive and up-to-date overview of FMs for large-scale ECG-FMs. First, we offer a brief background introduction to FMs. Then, we discuss the model architectures, pre-training methods, and adaptation approaches of ECG-FMs from a methodology perspective. Despite the promising opportunities of ECG-FMs, we also outline the challenges and potential future directions. Overall, this survey aims to provide researchers and practitioners with insights into the research of ECG-FMs on theoretical underpinnings, domain-specific applications, and avenues for future exploration.

arXiv:2410.20061

Deep Concept Identification for Generative Design

A generative design based on topology optimization provides diverse alternatives as entities in a computational model with a high design degree. However, as the diversity of the generated alternatives increases, the cognitive burden on designers to select the most appropriate alternatives also increases. Whereas the concept identification approach, which finds various categories of entities, is an effective means to structure alternatives, evaluation of their similarities is challenging due to shape diversity. To address this challenge, this study proposes a concept identification framework for generative design using deep learning (DL) techniques. One of the key abilities of DL is the automatic learning of different representations of a specific task. Deep concept identification finds various categories that provide insights into the mapping relationships between geometric properties and structural performance through representation learning using DL. The proposed framework generates diverse alternatives using a generative design technique, clusters the alternatives into several categories using a DL technique, and arranges these categories for design practice using a classification model. This study demonstrates its fundamental capabilities by implementing variational deep embedding, a generative and clustering model based on the DL paradigm, and logistic regression as a classification model. A simplified design problem of a two-dimensional bridge structure is applied as a case study to validate the proposed framework. Although designers are required to determine the viewing aspect level by setting the number of concepts, this implementation presents the identified concepts and their relationships in the form of a decision tree based on a specified level.

arXiv:2410.20969

BEVPose: Unveiling Scene Semantics through Pose-Guided Multi-Modal BEV Alignment

In the field of autonomous driving and mobile robotics, there has been a significant shift in the methods used to create Bird's Eye View (BEV) representations. This shift is characterised by using transformers and learning to fuse measurements from disparate vision sensors, mainly lidar and cameras, into a 2D planar ground-based representation. However, these learning-based methods for creating such maps often rely heavily on extensive annotated data, presenting notable challenges, particularly in diverse or non-urban environments where large-scale datasets are scarce. In this work, we present BEVPose, a framework that integrates BEV representations from camera and lidar data, using sensor pose as a guiding supervisory signal. This method notably reduces the dependence on costly annotated data. By leveraging pose information, we align and fuse multi-modal sensory inputs, facilitating the learning of latent BEV embeddings that capture both geometric and semantic aspects of the environment. Our pretraining approach demonstrates promising performance in BEV map segmentation tasks, outperforming fully-supervised state-of-the-art methods, while necessitating only a minimal amount of annotated data. This development not only confronts the challenge of data efficiency in BEV representation learning but also broadens the potential for such techniques in a variety of domains, including off-road and indoor environments.

arXiv:2410.19923공식 코드 · Papers with Code

Language Agents Meet Causality -- Bridging LLMs and Causal World Models

Large Language Models (LLMs) have recently shown great promise in planning and reasoning applications. These tasks demand robust systems, which arguably require a causal understanding of the environment. While LLMs can acquire and reflect common sense causal knowledge from their pretraining data, this information is often incomplete, incorrect, or inapplicable to a specific environment. In contrast, causal representation learning (CRL) focuses on identifying the underlying causal structure within a given environment. We propose a framework that integrates CRLs with LLMs to enable causally-aware reasoning and planning. This framework learns a causal world model, with causal variables linked to natural language expressions. This mapping provides LLMs with a flexible interface to process and generate descriptions of actions and states in text form. Effectively, the causal world model acts as a simulator that the LLM can query and interact with. We evaluate the framework on causal inference and planning tasks across temporal scales and environmental complexities. Our experiments demonstrate the effectiveness of the approach, with the causally-aware method outperforming LLM-based reasoners, especially for longer planning horizons.

arXiv:2410.20349

Idempotent Unsupervised Representation Learning for Skeleton-Based Action Recognition

Generative models, as a powerful technique for generation, also gradually become a critical tool for recognition tasks. However, in skeleton-based action recognition, the features obtained from existing pre-trained generative methods contain redundant information unrelated to recognition, which contradicts the nature of the skeleton's spatially sparse and temporally consistent properties, leading to undesirable performance. To address this challenge, we make efforts to bridge the gap in theory and methodology and propose a novel skeleton-based idempotent generative model (IGM) for unsupervised representation learning. More specifically, we first theoretically demonstrate the equivalence between generative models and maximum entropy coding, which demonstrates a potential route that makes the features of generative models more compact by introducing contrastive learning. To this end, we introduce the idempotency constraint to form a stronger consistency regularization in the feature space, to push the features only to maintain the critical information of motion semantics for the recognition task. Our extensive experiments on benchmark datasets, NTU RGB+D and PKUMMD, demonstrate the effectiveness of our proposed method. On the NTU 60 xsub dataset, we observe a performance improvement from 84.6$\%$ to 86.2$\%$. Furthermore, in zero-shot adaptation scenarios, our model demonstrates significant efficacy by achieving promising results in cases that were previously unrecognizable. Our project is available at \url{https://github.com/LanglandsLin/IGM}.

arXiv:2410.21043공식 코드 · Papers with Code

Disentangled and Self-Explainable Node Representation Learning

Node representations, or embeddings, are low-dimensional vectors that capture node properties, typically learned through unsupervised structural similarity objectives or supervised tasks. While recent efforts have focused on explaining graph model decisions, the interpretability of unsupervised node embeddings remains underexplored. To bridge this gap, we introduce DiSeNE (Disentangled and Self-Explainable Node Embedding), a framework that generates self-explainable embeddings in an unsupervised manner. Our method employs disentangled representation learning to produce dimension-wise interpretable embeddings, where each dimension is aligned with distinct topological structure of the graph. We formalize novel desiderata for disentangled and interpretable embeddings, which drive our new objective functions, optimizing simultaneously for both interpretability and disentanglement. Additionally, we propose several new metrics to evaluate representation quality and human interpretability. Extensive experiments across multiple benchmark datasets demonstrate the effectiveness of our approach.

arXiv:2410.20348공식 코드 · Papers with Code

UTSRMorph: A Unified Transformer and Superresolution Network for Unsupervised Medical Image Registration

Complicated image registration is a key issue in medical image analysis, and deep learning-based methods have achieved better results than traditional methods. The methods include ConvNet-based and Transformer-based methods. Although ConvNets can effectively utilize local information to reduce redundancy via small neighborhood convolution, the limited receptive field results in the inability to capture global dependencies. Transformers can establish long-distance dependencies via a self-attention mechanism; however, the intense calculation of the relationships among all tokens leads to high redundancy. We propose a novel unsupervised image registration method named the unified Transformer and superresolution (UTSRMorph) network, which can enhance feature representation learning in the encoder and generate detailed displacement fields in the decoder to overcome these problems. We first propose a fusion attention block to integrate the advantages of ConvNets and Transformers, which inserts a ConvNet-based channel attention module into a multihead self-attention module. The overlapping attention block, a novel cross-attention method, uses overlapping windows to obtain abundant correlations with match information of a pair of images. Then, the blocks are flexibly stacked into a new powerful encoder. The decoder generation process of a high-resolution deformation displacement field from low-resolution features is considered as a superresolution process. Specifically, the superresolution module was employed to replace interpolation upsampling, which can overcome feature degradation. UTSRMorph was compared to state-of-the-art registration methods in the 3D brain MR (OASIS, IXI) and MR-CT datasets. The qualitative and quantitative results indicate that UTSRMorph achieves relatively better performance. The code and datasets are publicly available at https://github.com/Runshi-Zhang/UTSRMorph.

arXiv:2410.19176

Perturbation-based Graph Active Learning for Weakly-Supervised Belief Representation Learning

This paper addresses the problem of optimizing the allocation of labeling resources for semi-supervised belief representation learning in social networks. The objective is to strategically identify valuable messages on social media graphs that are worth labeling within a constrained budget, ultimately maximizing the task's performance. Despite the progress in unsupervised or semi-supervised methods in advancing belief and ideology representation learning on social networks and the remarkable efficacy of graph learning techniques, the availability of high-quality curated labeled social data can greatly benefit and further improve performances. Consequently, allocating labeling efforts is a critical research problem in scenarios where labeling resources are limited. This paper proposes a graph data augmentation-inspired perturbation-based active learning strategy (PerbALGraph) that progressively selects messages for labeling according to an automatic estimator, obviating human guidance. This estimator is based on the principle that messages in the network that exhibit heightened sensitivity to structural features of the observational data indicate landmark quality that significantly influences semi-supervision processes. We design the estimator to be the prediction variance under a set of designed graph perturbations, which is model-agnostic and application-independent. Extensive experiment results demonstrate the effectiveness of the proposed strategy for belief representation learning tasks.

arXiv:2410.19174

Indication Finding: a novel use case for representation learning

Many therapies are effective in treating multiple diseases. We present an approach that leverages methods developed in natural language processing and real-world data to prioritize potential, new indications for a mechanism of action (MoA). We specifically use representation learning to generate embeddings of indications and prioritize them based on their proximity to the indications with the strongest available evidence for the MoA. We demonstrate the successful deployment of our approach for anti-IL-17A using embeddings generated with SPPMI and present an evaluation framework to determine the quality of indication finding results and the derived embeddings.

arXiv:2410.19722

Temporal Convolution-based Hybrid Model Approach with Representation Learning for Real-Time Acoustic Anomaly Detection

The early detection of potential failures in industrial machinery components is paramount for ensuring the reliability and safety of operations, thereby preserving Machine Condition Monitoring (MCM). This research addresses this imperative by introducing an innovative approach to Real-Time Acoustic Anomaly Detection. Our method combines semi-supervised temporal convolution with representation learning and a hybrid model strategy with Temporal Convolutional Networks (TCN) to handle various intricate anomaly patterns found in acoustic data effectively. The proposed model demonstrates superior performance compared to established research in the field, underscoring the effectiveness of this approach. Not only do we present quantitative evidence of its superiority, but we also employ visual representations, such as t-SNE plots, to further substantiate the model's efficacy.

arXiv:2410.19444

Balancing the Scales: Enhancing Fairness in Facial Expression Recognition with Latent Alignment

Automatically recognizing emotional intent using facial expression has been a thoroughly investigated topic in the realm of computer vision. Facial Expression Recognition (FER), being a supervised learning task, relies heavily on substantially large data exemplifying various socio-cultural demographic attributes. Over the past decade, several real-world in-the-wild FER datasets that have been proposed were collected through crowd-sourcing or web-scraping. However, most of these practically used datasets employ a manual annotation methodology for labeling emotional intent, which inherently propagates individual demographic biases. Moreover, these datasets also lack an equitable representation of various socio-cultural demographic groups, thereby inducing a class imbalance. Bias analysis and its mitigation have been investigated across multiple domains and problem settings, however, in the FER domain, this is a relatively lesser explored area. This work leverages representation learning based on latent spaces to mitigate bias in facial expression recognition systems, thereby enhancing a deep learning model's fairness and overall accuracy.

arXiv:2410.19283

ST-NeRP: Spatial-Temporal Neural Representation Learning with Prior Embedding for Patient-specific Imaging Study

During and after a course of therapy, imaging is routinely used to monitor the disease progression and assess the treatment responses. Despite of its significance, reliably capturing and predicting the spatial-temporal anatomic changes from a sequence of patient-specific image series presents a considerable challenge. Thus, the development of a computational framework becomes highly desirable for a multitude of practical applications. In this context, we propose a strategy of Spatial-Temporal Neural Representation learning with Prior embedding (ST-NeRP) for patient-specific imaging study. Our strategy involves leveraging an Implicit Neural Representation (INR) network to encode the image at the reference time point into a prior embedding. Subsequently, a spatial-temporally continuous deformation function is learned through another INR network. This network is trained using the whole patient-specific image sequence, enabling the prediction of deformation fields at various target time points. The efficacy of the ST-NeRP model is demonstrated through its application to diverse sequential image series, including 4D CT and longitudinal CT datasets within thoracic and abdominal imaging. The proposed ST-NeRP model exhibits substantial potential in enabling the monitoring of anatomical changes within a patient throughout the therapeutic journey.

arXiv:2410.19183

Can Self Supervision Rejuvenate Similarity-Based Link Prediction?

Although recent advancements in end-to-end learning-based link prediction (LP) methods have shown remarkable capabilities, the significance of traditional similarity-based LP methods persists in unsupervised scenarios where there are no known link labels. However, the selection of node features for similarity computation in similarity-based LP can be challenging. Less informative node features can result in suboptimal LP performance. To address these challenges, we integrate self-supervised graph learning techniques into similarity-based LP and propose a novel method: Self-Supervised Similarity-based LP (3SLP). 3SLP is suitable for the unsupervised condition of similarity-based LP without the assistance of known link labels. Specifically, 3SLP introduces a dual-view contrastive node representation learning (DCNRL) with crafted data augmentation and node representation learning. DCNRL is dedicated to developing more informative node representations, replacing the node attributes as inputs in the similarity-based LP backbone. Extensive experiments over benchmark datasets demonstrate the salient improvement of 3SLP, outperforming the baseline of traditional similarity-based LP by up to 21.2% (AUC).

arXiv:2410.19110

Bio2Token: All-atom tokenization of any biomolecular structure with Mamba

Efficient encoding and representation of large 3D molecular structures with high fidelity is critical for biomolecular design applications. Despite this, many representation learning approaches restrict themselves to modeling smaller systems or use coarse-grained approximations of the systems, for example modeling proteins at the resolution of amino acid residues rather than at the level of individual atoms. To address this, we develop quantized auto-encoders that learn atom-level tokenizations of complete proteins, RNA and small molecule structures with reconstruction accuracies well below 1 Angstrom. We demonstrate that a simple Mamba state space model architecture is efficient compared to an SE(3)-invariant IPA architecture, reaches competitive accuracies and can scale to systems with almost 100,000 atoms. The learned structure tokens of bio2token may serve as the input for all-atom generative models in the future.