检索结果-内蒙古大学图书馆

Span Attention for Entity-Consistent Task-Oriented Dialogue Response Generation

学校读者我要写书评

暂无评论

Span Attention for Entity-Consistent Task-Oriented Dialogue ...

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Chen, Jiale Dong, Xuelian Xie, Wenxiu Gong, Tao Wang, Fu Lee Hao, Tianyong School of Computer Science South China Normal University Guangzhou China Department of Computer Science City University of Hong Kong Hong Kong Google Inc. New York United States School of Science and Technology Hong Kong Metropolitan University Hong Kong

ISBN: (纸本)9798350368741

Task-oriented dialogue systems have recently gained increasing attention due to their capability of using natural language to fulfill specific user demands, such as restaurant reservation and hotel booking. Recent works directly model task-oriented dialogue response as a text generation task. However, these methods, utilizing generated response tokens as an attention query to obtain the vanilla attention distribution over an entire knowledge base, frequently lead to an entity inconsistency in final response generation. To tackle this problem, we propose a novel attention mechanism called span attention and a novel model named Span Attention GEnerator (SAGE). The span attention computes an attention score between a query vector and each knowledge record vector instead of computing a vanilla attention score among word vectors, which consisted of dialogue context and knowledge base. For effective training, we propose an attention constraint strategy that utilizes the entities appearing in response as pseudo-labels to supervise the training of the span attention. Experiments based on three publicly accessible datasets demonstrate that our model, utilizing the proposed mechanism, outperforms the state-of-the-art models with improvements of 8.94%, 2.29%, and 10.22% respectively in Entity F1. © 2025 IEEE.

关键词： knowledge retrieval response generation span attention spoken dialogue systems task-oriented

The Odyssey Journey: Top-Tier Medical Resource Seeking for Specialized Disorder in China 25

学校读者我要写书评

暂无评论

The Odyssey Journey: Top-Tier Medical Resource Seeking for S...

2025 CHI Conference on Human Factors in Computing Systems, CHI 2025

作者： Chan, Ka I Hu, Siying Wang, Yuntao Xu, Xuhai Lu, Zhicong Shi, Yuanchun Department of Computer Science and Technology Beijing National Research Center for Information Science and Technology Global Innovation Exchange (GIX) Institute Tsinghua University Beijing China Department of Computer Science City University of Hong Kong Hong Kong Key Laboratory of Pervasive Computing Ministry of Education Department of Computer Science and Technology Tsinghua University Beijing China School of Computer Science and Technology Qinghai University Xining China Google New York City NY United States Department of Computer Science George Mason University Fairfax VA United States Department of Computer Science and Technology Beijing National Research Center for Information Science and Technology Tsinghua University Beijing China Qinghai University Xining China

ISBN: (纸本)9798400713941

It is pivotal for patients to receive accurate health information, diagnoses, and timely treatments. However, in China, the significant imbalanced doctor-to-patient ratio intensifies the information and power asymmetries in doctor-patient relationships. Health information-seeking, which enables patients to collect information from sources beyond doctors, is a potential approach to mitigate these asymmetries. While HCI research predominantly focuses on common chronic conditions, our study focuses on specialized disorders, which are often familiar to specialists but not to general practitioners and the public. With Hemifacial Spasm (HFS) as an example, we aim to understand patients' health information and top-tier¹ medical resource seeking journeys in China. Through interviews with three neurosurgeons and 12 HFS patients from rural and urban areas, and applying Actor-Network Theory, we provide empirical insights into the roles, interactions, and workflows of various actors in the health information-seeking network. We also identified five strategies patients adopted to mitigate asymmetries and access top-tier medical resources, illustrating these strategies as subnetworks within the broader health information-seeking network and outlining their advantages and challenges. © 2025 Copyright held by the owner/author(s).

关键词： Actor-Network Theory Health Information Seeking Hemifacial Spasm

Robust Knowledge Distillation from RNN-T Models with Noisy Training Labels Using Full-Sum Loss 48

学校读者我要写书评

暂无评论

Robust Knowledge Distillation from RNN-T Models with Noisy T...

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Zeineldeen, Mohammad Audhkhasi, Kartik Baskar, Murali Karthick Ramabhadran, Bhuvana Rwth Aachen University Human Language Technology and Pattern Recognition Computer Science Department Aachen52074 Germany Google Llc New York United States

ISBN: (纸本)9781728163277

This work studies knowledge distillation (KD) and addresses its constraints for recurrent neural network transducer (RNN-T) models. In hard distillation, a teacher model transcribes large amounts of unlabelled speech to train a student model. Soft distillation is another popular KD method that distills the output logits of the teacher model. Due to the nature of RNN-T alignments, applying soft distillation between RNNT architectures having different posterior distributions is challenging. In addition, bad teachers having high word-error-rate (WER) reduce the efficacy of KD. We investigate how to effectively distill knowledge from variable quality ASR teachers, which has not been studied before to the best of our knowledge. We show that a sequence-level KD, full-sum distillation, outperforms other distillation methods for RNN-T models, especially for bad teachers. We also propose a variant of full-sum distillation that distills the sequence discriminative knowledge of the teacher leading to further improvement in WER. We conduct experiments on public datasets namely SpeechStew and LibriSpeech, and on in-house production data. © 2023 IEEE.

关键词： Recurrent neural networks

Unlocking Accuracy and Fairness in Differentially Private Image Classification

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Berrada, Leonard De, Soham Shen, Judy Hanwen Hayes, Jamie Stanforth, Robert Stutz, David Kohli, Pushmeet Smith, Samuel L. Balle, Borja Google DeepMind London United Kingdom Computer Science Department Stanford University Palo AltoCA United States

Privacy-preserving machine learning aims to train models on private data without leaking sensitive information. Differential privacy (DP) is considered the gold standard framework for privacy-preserving training, as it provides formal privacy guarantees. However, compared to their non-private counterparts, models trained with DP often have significantly reduced accuracy. Private classifiers are also believed to exhibit larger performance disparities across subpopulations, raising fairness concerns. The poor performance of classifiers trained with DP has prevented the widespread adoption of privacy preserving machine learning in industry. Here we show that pre-trained foundation models fine-tuned with DP can achieve similar accuracy to non-private classifiers, even in the presence of significant distribution shifts between pre-training data and downstream tasks. We achieve private accuracies within a few percent of the non-private state of the art across four datasets, including two medical imaging benchmarks. Furthermore, our private medical classifiers do not exhibit larger performance disparities across demographic groups than non-private models. This milestone to make DP training a practical and reliable technology has the potential to widely enable machine learning practitioners to train safely on sensitive datasets while protecting individuals’ privacy. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

Enhancing Value Estimation Policies by Post-Hoc Symmetry Exploitation in Motion Planning Tasks

学校读者我要写书评

暂无评论

Enhancing Value Estimation Policies by Post-Hoc Symmetry Exp...

IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

作者： Yazied Hasan Ariana M. Villegas-Suarez Evan C. Carter Aleksandra Faust Lydia Tapia Department - of Computer Science University of New Mexico USA DEVCOM Army Research Laboratory USA Google Deepmind USA

Motion planning tasks are often innately invariant to certain geometric transformations, or in other words, symmetric. This property, however, is not always reflected in learned policies that are trained on these tasks. Although this asymmetry can be addressed through data augmentation or additional training samples, doing so comes at a cost of increased training time. Instead of trying to remedy this issue during the learning process, we leverage this disparity during execution. We propose the symmetry exploitation policy, an augmentation in the post-hoc execution stage of RL policies. During the planning stage, we present the learned policy with an invariant, geometrically transformed version of the observation as an alternate perspective of the state. This allows the policy to produce multiple possible actions for a single state, and choose the action with the highest estimated value. Unlike other symmetry exploitation methods for learning solutions in motion planning, this method completely bypasses the need for additional training. We show the effect of the symmetry exploitation policy on DQN, A2C, and PPO policies, in three motion problems with different dimensions, observation types, and symmetries. The results show that by exploiting the symmetry of the task, a trained model achieves improved performance and better generalization, and can achieve comparable results to retraining, augmentation, or extended training, without incurring any additional training time. The efficacy is most prominent in more complex tasks, as 89 of the 100 models involved in the case study improve when using the method.

关键词：

Discovery and Expansion of New Domains within Diffusion Models

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Zhu, Ye Wu, Yu Xu, Duo Deng, Zhiwei Yan, Yan Russakovsky, Olga Department of Computer Science Princeton University United States School of Computer Science Wuhan University China Department of Astronomy University of Virginia United States Google DeepMind United Kingdom Department of Computer Science Illinois Institute of Technology United States

In this work, we study the generalization properties of diffusion models in a few-shot setup, introduce a novel tuning-free paradigm to synthesize the target out-of-domain (OOD) data, and demonstrate its advantages compared to existing methods in data-sparse scenarios with large domain gaps. Specifically, given a pre-trained model and a small set of images that are OOD relative to the model's training distribution, we explore whether the frozen model is able to generalize to this new domain. We begin by revealing that Denoising Diffusion Probabilistic Models (DDPMs) trained on single-domain images are already equipped with sufficient representation abilities to reconstruct arbitrary images from the inverted latent encoding following bi-directional deterministic diffusion and denoising trajectories. We then demonstrate through both theoretical and empirical perspectives that the OOD images establish Gaussian priors in latent spaces of the given model, and the inverted latent modes are separable from their initial training domain. We then introduce our novel tuning-free paradigm to synthesize new images of the target unseen domain by discovering qualified OOD latent encodings in the inverted noisy spaces. This is fundamentally different from the current paradigm that seeks to modify the denoising trajectory to achieve the same goal by tuning the model parameters. Extensive cross-model and domain experiments show that our proposed method can expand the latent space and generate unseen images via frozen DDPMs without impairing the quality of generation of their original domain. We also showcase a practical application of our proposed heuristic approach in dramatically different domains using astrophysical data, revealing the great potential of such a generalization paradigm in data spare fields such as scientific explorations. © 2023, CC BY-NC-SA.

关键词： Diffusion

Learning Energy-Based Prior Model with Diffusion-Amortized MCMC

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Yu, Peiyu Zhu, Yaxuan Xie, Sirui Ma, Xiaojian Gao, Ruiqi Zhu, Song-Chun Wu, Ying Nian UCLA Department of Statistics United States UCLA Department of Computer Science United States Google DeepMind United Kingdom China

Latent space Energy-Based Models (EBMs), also known as energy-based priors, have drawn growing interests in the field of generative modeling due to its flexibility in the formulation and strong modeling power of the latent space. However, the common practice of learning latent space EBMs with non-convergent short-run MCMC for prior and posterior sampling is hindering the model from further progress;the degenerate MCMC sampling quality in practice often leads to degraded generation quality and instability in training, especially with highly multimodal and/or high-dimensional target distributions. To remedy this sampling issue, in this paper we introduce a simple but effective diffusion-based amortization method for long-run MCMC sampling and develop a novel learning algorithm for the latent space EBM based on it. We provide theoretical evidence that the learned amortization of MCMC is a valid long-run MCMC sampler. Experiments on several image modeling benchmark datasets demonstrate the superior performance of our method compared with strong counterparts1 © 2023, CC BY-SA.

关键词： Benchmarking

MULAN: A Study of Fact Mutability in Language Models

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Fierro, Constanza Garneau, Nicolas Bugliarello, Emanuele Kementchedjhieva, Yova Søgaard, Anders Department of Computer Science University of Copenhagen Denmark Mohamed bin Zayed University of Artificial Intelligence United Arab Emirates Google DeepMind United Kingdom

Facts are subject to contingencies and can be true or false in different circumstances. One such contingency is time, wherein some facts mutate over a given period, e.g., the president of a country or the winner of a championship. Trustworthy language models ideally identify mutable facts as such and process them accordingly. We create MULAN, a benchmark for evaluating the ability of English language models to anticipate time-contingency, covering both 1:1 and 1:N relations. We hypothesize that mutable facts are encoded differently than immutable ones, hence being easier to update. In a detailed evaluation of six popular large language models, we consistently find differences in the LLMs' confidence, representations, and update behavior, depending on the mutability of a fact. Our findings should inform future work on the injection of and induction of time-contingent knowledge to/from LLMs. © 2024, CC BY.

关键词： Computational linguistics

Frequency and Generalization of Periodic Activation Functions in Reinforcement Learning

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Mavor-Parker, Augustine N. Sargent, Matthew J. Barry, Caswell Griffin, Lewis D. Lyle, Clare Department of Computer Science University College London United Kingdom Department of Cell and Developmental Biology University College London United Kingdom Google Deepmind United Kingdom

Periodic activation functions, often referred to as learned Fourier features have been demonstrated to improve the sample efficiency and stability of deep RL algorithms. Ostensibly incompatible hypotheses have been made about the source of these improvements. One is that periodic activations learn low frequency representations and as a result avoid overfitting to bootstrapped targets. Another is that periodic activations learn high frequency representations that are more expressive, allowing networks to quickly fit complex value functions. We analyze these claims empirically, finding that periodic activation function architectures consistently converge to high frequency representations regardless of their initialization frequencies. As a result, we also find that while periodic activation functions improve sample efficiency, they exhibit worse generalization on states with added observation noise—especially when compared to otherwise identical networks with ReLU activation functions. Finally, we show that weight decay regularization is able to partially offset the overfitting of periodic activation functions, delivering value functions that learn quickly while also generalizing. © 2024, CC BY.

关键词： Deep reinforcement learning