检索结果-内蒙古大学图书馆

Discovering Chunks in Neural Embeddings for Interpretability

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wu, Shuchen Alaniz, Stephan Schulz, Eric Akata, Zeynep Explainable Machine Learning Lab Helmholtz Munich Germany Max Planck Institute for Biological Cybernetics Germany Munich Center for Machine Learning Germany Department of Computer Science Technical University of Munich Germany Institute for Human-Centered AI Helmholtz Munich Germany

Understanding neural networks is challenging due to their high-dimensional, interacting components. Inspired by human cognition, which processes complex sensory data by chunking it into recurring entities, we propose leveraging this principle to interpret artificial neural population activities. Biological and artificial intelligence share the challenge of learning from structured, naturalistic data, and we hypothesize that the cognitive mechanism of chunking can provide insights into artificial systems. We first demonstrate this concept in recurrent neural networks (RNNs) trained on artificial sequences with imposed regularities, observing that their hidden states reflect these patterns, which can be extracted as a dictionary of chunks that influence network responses. Extending this to large language models (LLMs) like LLaMA, we identify similar recurring embedding states corresponding to concepts in the input, with perturbations to these states activating or inhibiting the associated concepts. By exploring methods to extract dictionaries of identifiable chunks across neural embeddings of varying complexity, our findings introduce a new framework for interpreting neural networks, framing their population activity as structured reflections of the data they process. © 2025, CC BY.

关键词： Network embeddings

Predicting Price Direction of Bitcoin based on Hybrid Model of LSTM and Dense Neural Network Approach 4

学校读者我要写书评

暂无评论

Predicting Price Direction of Bitcoin based on Hybrid Model ...

4th International Conference on Electronics and Sustainable Communication Systems, ICESC 2023

作者： Karahyla, Jai Kishan Sharma, Neelam Chamoli, Sushant Shirgire, Anil Kant, Ravi Chauhan, Amit Pgcc Hospital Management Professor and Head Respiratory Medicine Mmimsr Mm Deemed University Human Rights HaryanaMullana Ambala India Lovely Professional University Physical Education Punjab Phagwara India Graphic Era Hill University Department of Computer Science and Engineering Dehradun India Dr D y Patil Institute of Technology Department of Civil Engineering MaharashtraPimpri Pune India Shoolini University Cse Department Himachal Pradesh Solan India Department of Life Sciences Karnataka Bengaluru India

ISBN: (纸本)9798350300093

Bitcoin is a rapidly growing but extremely risky cryptocurrency. It marks a watershed moment in the history of cash. These days, digital currency is preferred to actual money. Bitcoin has decentralized authority and placed it in the hands of its users. Many people are joining the largest and most well-known Bitcoin mining pools as the risk of working alone is too great. In order to enhance their chances of creating the next block in the Bitcoins blockchain and decrease the mining reward volatility, users can band together to form Bitcoin pools. This tendency toward consolidation may also be seen in the rise of large-scale mining farms equipped with powerful mining resources and speedy processing capability. Because of the risk of a 51% assault, this pattern shows that Bitcoin's pure, decentralized protocol is moving toward greater centralization in its distribution network. Not to be overlooked is the resulting centralization of the bitcoin network as a result of cloud wallets making it simple for new users to join. Because of the easily hackable nature of Bitcoin technologies, this could lead to a wide range of security vulnerabilities. The proposed approach uses normalization and filling missing values in preprocessing, PCA for feature Extraction and finally training the model using LSTM-DNN Models. The proposed approach outperforms other two models such as CNN and DNN. © 2023 IEEE.

关键词： Bitcoin

A hybrid deep learning architecture using 3D CNNs and GRUs for human action recognition

学校读者我要写书评

暂无评论

International Journal of Engineering, Transactions B: Applications 2020年第5期33卷 959-965页

作者： Savadi Hosseini, M. Ghaderi, F. Human-Computer Interaction lab Department of Electrical and Computer Engineering Tarbiat Modares University Tehran Iran

Video contents have variations in temporal and spatial dimensions, and recognizing human actions requires considering the changes in both directions. To this end, convolutional neural networks (CNNs) and recurrent neural networks (RNNs) and their combinations have been used to tackle the video dynamics. However, a hybrid architecture usually results in a more complex model and hence a greater number of parameters to be optimized. In this study, we propose to use a stack of gated recurrent unit (GRU) layers on top of a two-stream inflated convolutional neural network. Raw frames and optical flow of the video are processed in the first and second streams, respectively. We first segment the video frames in order to be able to track the video contents in more details and by using 3D CNNs extract spatial-temporal features, called local features. We then import the sequence of local features to the GRU network, and use a weighted averaging operator to aggregate the outcome of the two processing flows, called global features. The evaluations confirm acceptable results for the two HMDB51 and UCF101 datasets. The proposed method resulted in a 1.6% improvement in the classification accuracy of the HMDB51 challenging dataset compared to the best reported results. © 2020 Materials and Energy Research Center. All rights reserved.

关键词： Convolutional neural networks

A Self-supervised Framework for Improved Data-Driven Monitoring of Stress via Multi-modal Passive Sensing

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Fazeli, Shayan Levine, Lionel Beikzadeh, Mehrab Mirzasoleiman, Baharan Zadeh, Bita Peris, Tara Sarrafzadeh, Majid Computer Science Department Ucla Los Angeles United States Department of Psychology Chapman University Los Angeles United States The Jane and Terry Semel Institute for Neuroscience and Human Behavior Ucla Los Angeles United States

Recent advances in remote health monitoring systems have significantly benefited patients and played a crucial role in improving their quality of life. However, while physiological health-focused solutions have demonstrated increasing success and maturity, mental health-focused applications have seen comparatively limited success in spite of the fact that stress and anxiety disorders are among the most common issues people deal with in their daily lives. In the hopes of furthering progress in this domain through the development of a more robust analytic framework for the measurement of indicators of mental health, we propose a multi-modal semi-supervised framework for tracking physiological precursors of the stress response. Our methodology enables utilizing multi-modal data of differing domains and resolutions from wearable devices and leveraging them to map short-term episodes to semantically efficient embeddings for a given task. Additionally, we leverage an inter-modality contrastive objective, with the advantages of rendering our framework both modular and scalable. The focus on optimizing both local and global aspects of our embeddings via a hierarchical structure renders transferring knowledge and compatibility with other devices easier to achieve. In our pipeline, a task-specific pooling based on an attention mechanism, which estimates the contribution of each modality on an instance level, computes the final embeddings for observations. This additionally provides a thorough diagnostic insight into the data characteristics and highlights the importance of signals in the broader view of predicting episodes annotated per mental health status. We perform training experiments using a corpus of realworld data on perceived stress, and our results demonstrate the efficacy of the proposed approach in performance improvements. © 2023, CC BY-NC-SA.

关键词： Embeddings

Parts2Words: Learning Joint Embedding of Point Clouds and Texts by Bidirectional Matching Between Parts and Words

学校读者我要写书评

暂无评论

Parts2Words: Learning Joint Embedding of Point Clouds and Te...

Conference on computer Vision and Pattern Recognition (CVPR)

作者： Chuan Tang Xi Yang Bojian Wu Zhizhong Han Yi Chang School of Artificial Intelligence Jilin University China MoE Engineering Research Center of Knowledge-Driven Human-Machine Intelligence China Zhejiang University The Department of Computer Science Wayne State University USA

Shape-Text matching is an important task of high-level shape understanding. Current methods mainly represent a 3D shape as multiple 2D rendered views, which obviously can not be understood well due to the structural ambiguity caused by self-occlusion in the limited number of views. To resolve this issue, we directly represent 3D shapes as point clouds, and propose to learn joint embedding of point clouds and texts by bidirectional matching between parts from shapes and words from texts. Specifically, we first segment the point clouds into parts, and then leverage optimal transport method to match parts and words in an optimized feature space, where each part is represented by aggregating features of all points within it and each word is abstracted by its contextual information. We optimize the feature space in order to enlarge the similarities between the paired training samples, while simultaneously maximizing the margin between the unpaired ones. Experiments demonstrate that our method achieves a significant improvement in accuracy over the SOTAs on multi-modal retrieval tasks under the Text2Shape dataset. Codes are available at here.

关键词：

Plug-and-Play Model-Agnostic Counterfactual Policy Synthesis for Deep Reinforcement Learning based Recommendation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wang, Siyu Chen, Xiaocong McAuley, Julian Cripps, Sally Yao, Lina The School of Computer Science and Engineering University of New South Wales SydneyNSW2052 Australia Data61 CSIRO Australia Human Technology Institute University of Technology Sydney Australia Computer Science Department UCSD United States

Recent advances in recommender systems have proved the potential of Reinforcement Learning (RL) to handle the dynamic evolution processes between users and recommender systems. However, learning to train an optimal RL agent is generally impractical with commonly sparse user feedback data in the context of recommender systems. To circumvent the lack of interaction of current RL-based recommender systems, we propose to learn a general Model-Agnostic Counterfactual Synthesis (MACS) Policy for counterfactual user interaction data augmentation. The counterfactual synthesis policy aims to synthesise counterfactual states while preserving significant information in the original state relevant to the user's interests, building upon two different training approaches we designed: learning with expert demonstrations and joint training. As a result, the synthesis of each counterfactual data is based on the current recommendation agent's interaction with the environment to adapt to users' dynamic interests. We integrate the proposed policy Deep Deterministic Policy Gradient (DDPG), Soft Actor Critic (SAC) and Twin Delayed DDPG in an adaptive pipeline with a recommendation agent that can generate counterfactual data to improve the performance of recommendation. The empirical results on both online simulation and offline datasets demonstrate the effectiveness and generalisation of our counterfactual synthesis policy and verify that it improves the performance of RL recommendation agents. Copyright © 2022, The Authors. All rights reserved.

关键词： Reinforcement learning

How Deep is Love in LLMs’ Hearts? Exploring Semantic Size in human-like Cognition

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Yao, Yao Yang, Yifei Ma, Xinbei Yang, Dongjie Zhang, Zhuosheng Li, Zuchao Zhao, Hai Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai China Shanghai Key Laboratory of Trusted Data Circulation and Governance in Web3 Shanghai China Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai Jiao Tong University Shanghai China School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai China National Engineering Research Center for Multimedia Software School of Computer Science Wuhan University Wuhan University Wuhan430072 China

How human cognitive abilities are formed has long captivated researchers. However, a significant challenge lies in developing meaningful methods to measure these complex processes. With the advent of large language models (LLMs), which now rival human capabilities in various domains, we are presented with a unique testbed to investigate human cognition through a new lens. Among the many facets of cognition, one particularly crucial aspect is the concept of semantic size—the perceived magnitude of both abstract and concrete words or concepts. This study seeks to investigate whether LLMs exhibit similar tendencies in understanding semantic size, thereby providing insights into the underlying mechanisms of human cognition. We begin by exploring metaphorical reasoning, comparing how LLMs and humans associate abstract words with concrete objects of varying sizes. Next, we examine LLMs’ internal representations to evaluate their alignment with human cognitive processes. Our findings reveal that multi-modal training is crucial for LLMs to achieve more human-like understanding, suggesting that real-world, multi-modal experiences are similarly vital for human cognitive development. Lastly, we examine whether LLMs are influenced by attention-grabbing headlines with larger semantic sizes in a real-world web shopping scenario. The results show that multi-modal LLMs are more emotionally engaged in decision-making, but this also introduces potential biases, such as the risk of manipulation through clickbait headlines. Ultimately, this study offers a novel perspective on how LLMs interpret and internalize language, from the smallest concrete objects to the most profound abstract concepts like ‘love’. The insights gained not only improve our understanding of LLMs but also provide new avenues for exploring the cognitive abilities that define human intelligence. Our code and dataset is publicly available at https://***/Zoeyyao27/LLM-Semantic-Size-Understanding. © 2025, CC BY.

关键词： Semantics

Deep Generative Models for Constructing Synthetic Populations in Public Health Applications

学校读者我要写书评

暂无评论

Deep Generative Models for Constructing Synthetic Population...

Applied Imagery Pattern Recognition Workshop (AIPR)

作者： Nathan Farmer Arash Mahyari Ashok Sirinivasan Department of Computer Science University of West Florida Pensacola FL USA Florida Institute For Human and Machine Cognition (IHMC) Pensacola FL USA Department of Intelligent Systems and Robotics University of West Florida Pensacola FL USA

There is a growing acknowledgment of the significance of social determinants of health in shaping fair public health policies. These determinants, reflected in emerging data streams like mobility and social media data, are increasingly integral to public health models. However, privacy concerns impede broad access to sensitive data, because even non-identifiable data are susceptible to deductive disclosure. To address this, synthetic populations trained on such data emerge as a privacy-conscious solution, offering the added benefit of exploring various "what-if" scenarios. This paper delves into the application of Generative Adversarial Networks (GANs) to craft synthetic populations that tackle privacy issues linked to social determinants of health data. Utilizing the 2019 PUMS dataset for Florida, the study trains the GAN to produce diverse and adaptable synthetic populations. An autoencoder manages the mix of continuous and categorical variables, ensuring the generated data aligns with categorical ranges. The results, assessed by comparing marginal distributions and employing tSNE, underscore the GAN's efficacy in balancing privacy and data utility. The significance of this work lies in identifying the promise of GANs in generating synthetic populations to advance health research.

关键词：

Learning Time-Series Images of Niacin Skin-Flushing Test for the Diagnosis of Schizophrenia and Affective Disorder

学校读者我要写书评

暂无评论

Learning Time-Series Images of Niacin Skin-Flushing Test for...

2022 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2022

作者： Huang, Yuyang Dong, Shiwen Wang, Dandan Wan, Chunling Yang, Yang Shanghai Jiao Tong University Department of Computer Science and Engineering Shanghai China Shanghai Jiao Tong University Department of Electronic Engineering Shanghai China Bio-X Institutes Key Laboratory for the Genetics of Developmental and Neuropsychiatric Disorders Ministry of Education Shanghai Jiao Tong University Shanghai200030 China Key Lab. of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Shanghai200240 China

ISBN: (纸本)9781665468190

Niacin skin-flushing test is a promising method for fast and objective diagnosis of schizophrenia and affective disorder. Compared to healthy controls, the patients have attenuated flushing extent and reduced flushing rate. The traditional analysis of niacin test relies on manually evaluated flushing extent, while flushing patterns sh own in the sk in test images have not yet been considered. To exploit the potential of raw images for diagnosis and avoid subjective and time-consuming estimation of the flushing levels, we propose a CNN-LSTM hybrid model to learn the time-series of images captured in the niacin skin-flushing test. Moreover, an attention layer is introduced to enhance the model performance and interpretability. We collect a data set of 796 participants, and the experimental results show that the new model outperforms the traditional 4-point scaled measurement and conventional deep learning methods by large margins. Besides, the attention scores yielded by the model reveals key time points during the flushing reaction. © 2022 IEEE.

关键词： Time series