Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.09657
Jiacheng Yao, Weihong Xu, Zhaohui Yang, Xiaohu You, M. Bennis, H. V. Poor
In this paper, we quantitatively compare these two effective communication schemes, i.e., digital and analog ones, for wireless federated learning (FL) over resource-constrained networks, highlighting their essential differences as well as their respective application scenarios. We first examine both digital and analog transmission methods, together with a unified and fair comparison scheme under practical constraints. A universal convergence analysis under various imperfections is established for FL performance evaluation in wireless networks. These analytical results reveal that the fundamental difference between the two paradigms lies in whether communication and computation are jointly designed or not. The digital schemes decouple the communication design from specific FL tasks, making it difficult to support simultaneous uplink transmission of massive devices with limited bandwidth. In contrast, the analog communication allows over-the-air computation (AirComp), thus achieving efficient spectrum utilization. However, computation-oriented analog transmission reduces power efficiency, and its performance is sensitive to computational errors. Finally, numerical simulations are conducted to verify these theoretical observations.
{"title":"Digital versus Analog Transmissions for Federated Learning over Wireless Networks","authors":"Jiacheng Yao, Weihong Xu, Zhaohui Yang, Xiaohu You, M. Bennis, H. V. Poor","doi":"10.48550/arXiv.2402.09657","DOIUrl":"https://doi.org/10.48550/arXiv.2402.09657","url":null,"abstract":"In this paper, we quantitatively compare these two effective communication schemes, i.e., digital and analog ones, for wireless federated learning (FL) over resource-constrained networks, highlighting their essential differences as well as their respective application scenarios. We first examine both digital and analog transmission methods, together with a unified and fair comparison scheme under practical constraints. A universal convergence analysis under various imperfections is established for FL performance evaluation in wireless networks. These analytical results reveal that the fundamental difference between the two paradigms lies in whether communication and computation are jointly designed or not. The digital schemes decouple the communication design from specific FL tasks, making it difficult to support simultaneous uplink transmission of massive devices with limited bandwidth. In contrast, the analog communication allows over-the-air computation (AirComp), thus achieving efficient spectrum utilization. However, computation-oriented analog transmission reduces power efficiency, and its performance is sensitive to computational errors. Finally, numerical simulations are conducted to verify these theoretical observations.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139963124","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.10065
Achraf Azize, Debabrota Basu
We study the per-datum Membership Inference Attacks (MIAs), where an attacker aims to infer whether a fixed target datum has been included in the input dataset of an algorithm and thus, violates privacy. First, we define the membership leakage of a datum as the advantage of the optimal adversary targeting to identify it. Then, we quantify the per-datum membership leakage for the empirical mean, and show that it depends on the Mahalanobis distance between the target datum and the data-generating distribution. We further assess the effect of two privacy defences, i.e. adding Gaussian noise and sub-sampling. We quantify exactly how both of them decrease the per-datum membership leakage. Our analysis builds on a novel proof technique that combines an Edgeworth expansion of the likelihood ratio test and a Lindeberg-Feller central limit theorem. Our analysis connects the existing likelihood ratio and scalar product attacks, and also justifies different canary selection strategies used in the privacy auditing literature. Finally, our experiments demonstrate the impacts of the leakage score, the sub-sampling ratio and the noise scale on the per-datum membership leakage as indicated by the theory.
{"title":"How Much Does Each Datapoint Leak Your Privacy? Quantifying the Per-datum Membership Leakage","authors":"Achraf Azize, Debabrota Basu","doi":"10.48550/arXiv.2402.10065","DOIUrl":"https://doi.org/10.48550/arXiv.2402.10065","url":null,"abstract":"We study the per-datum Membership Inference Attacks (MIAs), where an attacker aims to infer whether a fixed target datum has been included in the input dataset of an algorithm and thus, violates privacy. First, we define the membership leakage of a datum as the advantage of the optimal adversary targeting to identify it. Then, we quantify the per-datum membership leakage for the empirical mean, and show that it depends on the Mahalanobis distance between the target datum and the data-generating distribution. We further assess the effect of two privacy defences, i.e. adding Gaussian noise and sub-sampling. We quantify exactly how both of them decrease the per-datum membership leakage. Our analysis builds on a novel proof technique that combines an Edgeworth expansion of the likelihood ratio test and a Lindeberg-Feller central limit theorem. Our analysis connects the existing likelihood ratio and scalar product attacks, and also justifies different canary selection strategies used in the privacy auditing literature. Finally, our experiments demonstrate the impacts of the leakage score, the sub-sampling ratio and the noise scale on the per-datum membership leakage as indicated by the theory.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139963341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.09683
Zeya Chen, Ruth Schmidt
Designing seamless, frictionless user experiences has long been a dominant trend in both applied behavioral science and artificial intelligence (AI), in which the goal of making desirable actions easy and efficient informs efforts to minimize friction in user experiences. However, in some settings, friction can be genuinely beneficial, such as the insertion of deliberate delays to increase reflection, preventing individuals from resorting to automatic or biased behaviors, and enhancing opportunities for unexpected discoveries. More recently, the popularization and availability of AI on a widespread scale has only increased the need to examine how friction can help or hinder users of AI; it also suggests a need to consider how positive friction can benefit AI practitioners, both during development processes (e.g., working with diverse teams) and to inform how AI is designed into offerings. This paper first proposes a"positive friction"model that can help characterize how friction is currently beneficial in user and developer experiences with AI, diagnose the potential need for friction where it may not yet exist in these contexts, and inform how positive friction can be used to generate solutions, especially as advances in AI continue to be progress and new opportunities emerge. It then explores this model in the context of AI users and developers by proposing the value of taking a hybrid"AI+human"lens, and concludes by suggesting questions for further exploration.
{"title":"Exploring a Behavioral Model of \"Positive Friction\" in Human-AI Interaction","authors":"Zeya Chen, Ruth Schmidt","doi":"10.48550/arXiv.2402.09683","DOIUrl":"https://doi.org/10.48550/arXiv.2402.09683","url":null,"abstract":"Designing seamless, frictionless user experiences has long been a dominant trend in both applied behavioral science and artificial intelligence (AI), in which the goal of making desirable actions easy and efficient informs efforts to minimize friction in user experiences. However, in some settings, friction can be genuinely beneficial, such as the insertion of deliberate delays to increase reflection, preventing individuals from resorting to automatic or biased behaviors, and enhancing opportunities for unexpected discoveries. More recently, the popularization and availability of AI on a widespread scale has only increased the need to examine how friction can help or hinder users of AI; it also suggests a need to consider how positive friction can benefit AI practitioners, both during development processes (e.g., working with diverse teams) and to inform how AI is designed into offerings. This paper first proposes a\"positive friction\"model that can help characterize how friction is currently beneficial in user and developer experiences with AI, diagnose the potential need for friction where it may not yet exist in these contexts, and inform how positive friction can be used to generate solutions, especially as advances in AI continue to be progress and new opportunities emerge. It then explores this model in the context of AI users and developers by proposing the value of taking a hybrid\"AI+human\"lens, and concludes by suggesting questions for further exploration.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139962168","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.09870
P. Koelewijn, Siep Weiland, Roland T'oth
This paper considers the equilibrium-free stability and performance analysis of discrete-time nonlinear systems. We consider two types of equilibrium-free notions. Namely, the universal shifted concept, which considers stability and performance w.r.t. all equilibrium points of the system, and the incremental concept, which considers stability and performance between trajectories of the system. In this paper, we show how universal shifted stability and performance of discrete-time systems can be analyzed by making use of the time-difference dynamics. Moreover, we extend the existing results for incremental dissipativity for discrete-time systems based on dissipativity analysis of the differential dynamics to more general state-dependent storage functions for less conservative results. Finally, we show how both these equilibrium-free notions can be cast as a convex analysis problem by making use of the linear parameter-varying framework, which is also demonstrated by means of an example.
{"title":"Convex Equilibrium-Free Stability and Performance Analysis of Discrete-Time Nonlinear Systems","authors":"P. Koelewijn, Siep Weiland, Roland T'oth","doi":"10.48550/arXiv.2402.09870","DOIUrl":"https://doi.org/10.48550/arXiv.2402.09870","url":null,"abstract":"This paper considers the equilibrium-free stability and performance analysis of discrete-time nonlinear systems. We consider two types of equilibrium-free notions. Namely, the universal shifted concept, which considers stability and performance w.r.t. all equilibrium points of the system, and the incremental concept, which considers stability and performance between trajectories of the system. In this paper, we show how universal shifted stability and performance of discrete-time systems can be analyzed by making use of the time-difference dynamics. Moreover, we extend the existing results for incremental dissipativity for discrete-time systems based on dissipativity analysis of the differential dynamics to more general state-dependent storage functions for less conservative results. Finally, we show how both these equilibrium-free notions can be cast as a convex analysis problem by making use of the linear parameter-varying framework, which is also demonstrated by means of an example.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139962363","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.09790
Chiara Garavelli, A. Aldieri, M. Palanca, Luca Patruno, M. Viceconti
Vertebral fractures prediction in clinics lacks of accuracy. The most used scores have limitations in distinguishing between subjects at risk or not. Finite element (FE) models generated from computed tomography (CT) of these patients may improve the predictive capability. Many models have already been proposed but the most of them considered the single vertebral body, excluding from the analysis the role of the inter-vertebral discs in the distribution of the load through the spine. Multi-vertebral models instead allow to examine more complex boundary condition. However, CT scans do not provide subject-specif information about the material properties of the disc. Consequently, the goal of the study was to validate a multi-vertebral FE model with subject specific modelling of the vertebral bone and population-based properties assigned to the disc, idealizing them with a linear isotropic material. Boundary condition were assigned in order to reproduce an experimental test performed on the same specimen and recorded using digital image correlation technique (DIC). FE and DIC strains on the vertebral surfaces are compared point-wise. Young's modulus values in the range 25-30 MPa allowed to achieve a comparable order of magnitude between experimental and computational data. However, the two distribution remained strongly different. To conclude, subject-specific material properties need to be assigned also to the discs as well as to the vertebrae to achieve acceptable accuracy in the assessment of the fracture risk.
临床上对椎体骨折的预测缺乏准确性。最常用的评分在区分受试者是否存在风险方面存在局限性。根据这些患者的计算机断层扫描(CT)生成的有限元(FE)模型可以提高预测能力。目前已经提出了许多模型,但其中大多数都只考虑了单个椎体,将椎间盘在脊柱负荷分布中的作用排除在分析之外。多椎体模型则可以研究更复杂的边界条件。然而,CT 扫描无法提供有关椎间盘材料特性的特定信息。因此,本研究的目标是验证一个多椎体 FE 模型,该模型具有针对特定受试者的椎骨建模和基于人群的椎间盘属性,将其理想化为线性各向同性材料。设定边界条件是为了重现在同一试样上进行的实验测试,该测试使用数字图像相关技术(DIC)记录。对椎体表面的 FE 应变和 DIC 应变进行了点对点比较。杨氏模量值在 25-30 兆帕之间,因此实验数据和计算数据的数量级相当。然而,两者的分布仍然存在很大差异。总之,需要为椎间盘和椎体分配特定的材料属性,以便在评估骨折风险时达到可接受的准确性。
{"title":"Multi-vertebral CT-based FE models implementing linear isotropic population-based material properties for the intervertebral discs cannot accurately predict strains","authors":"Chiara Garavelli, A. Aldieri, M. Palanca, Luca Patruno, M. Viceconti","doi":"10.48550/arXiv.2402.09790","DOIUrl":"https://doi.org/10.48550/arXiv.2402.09790","url":null,"abstract":"Vertebral fractures prediction in clinics lacks of accuracy. The most used scores have limitations in distinguishing between subjects at risk or not. Finite element (FE) models generated from computed tomography (CT) of these patients may improve the predictive capability. Many models have already been proposed but the most of them considered the single vertebral body, excluding from the analysis the role of the inter-vertebral discs in the distribution of the load through the spine. Multi-vertebral models instead allow to examine more complex boundary condition. However, CT scans do not provide subject-specif information about the material properties of the disc. Consequently, the goal of the study was to validate a multi-vertebral FE model with subject specific modelling of the vertebral bone and population-based properties assigned to the disc, idealizing them with a linear isotropic material. Boundary condition were assigned in order to reproduce an experimental test performed on the same specimen and recorded using digital image correlation technique (DIC). FE and DIC strains on the vertebral surfaces are compared point-wise. Young's modulus values in the range 25-30 MPa allowed to achieve a comparable order of magnitude between experimental and computational data. However, the two distribution remained strongly different. To conclude, subject-specific material properties need to be assigned also to the discs as well as to the vertebrae to achieve acceptable accuracy in the assessment of the fracture risk.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139962504","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.09731
Jianming Xian
Deep convolutional neural networks (CNNs) based approaches have achieved great performance in video matting. Many of these methods can produce accurate alpha estimation for the target body but typically yield fuzzy or incorrect target edges. This is usually caused by the following reasons: 1) The current methods always treat the target body and edge indiscriminately; 2) Target body dominates the whole target with only a tiny proportion target edge. For the first problem, we propose a CNN-based module that separately optimizes the matting target body and edge (SOBE). And on this basis, we introduce a real-time, trimap-free video matting method via progressively optimizing the matting target body and edge (POBEVM) that is much lighter than previous approaches and achieves significant improvements in the predicted target edge. For the second problem, we propose an Edge-L1-Loss (ELL) function that enforces our network on the matting target edge. Experiments demonstrate our method outperforms prior trimap-free matting methods on both Distinctions-646 (D646) and VideoMatte240K(VM) dataset, especially in edge optimization.
{"title":"POBEVM: Real-time Video Matting via Progressively Optimize the Target Body and Edge","authors":"Jianming Xian","doi":"10.48550/arXiv.2402.09731","DOIUrl":"https://doi.org/10.48550/arXiv.2402.09731","url":null,"abstract":"Deep convolutional neural networks (CNNs) based approaches have achieved great performance in video matting. Many of these methods can produce accurate alpha estimation for the target body but typically yield fuzzy or incorrect target edges. This is usually caused by the following reasons: 1) The current methods always treat the target body and edge indiscriminately; 2) Target body dominates the whole target with only a tiny proportion target edge. For the first problem, we propose a CNN-based module that separately optimizes the matting target body and edge (SOBE). And on this basis, we introduce a real-time, trimap-free video matting method via progressively optimizing the matting target body and edge (POBEVM) that is much lighter than previous approaches and achieves significant improvements in the predicted target edge. For the second problem, we propose an Edge-L1-Loss (ELL) function that enforces our network on the matting target edge. Experiments demonstrate our method outperforms prior trimap-free matting methods on both Distinctions-646 (D646) and VideoMatte240K(VM) dataset, especially in edge optimization.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139962602","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.09734
Paulo Garcia
Ensuring artificial intelligence behaves in such a way that is aligned with human values is commonly referred to as the alignment challenge. Prior work has shown that rational agents, behaving in such a way that maximizes a utility function, will inevitably behave in such a way that is not aligned with human values, especially as their level of intelligence goes up. Prior work has also shown that there is no"one true utility function"; solutions must include a more holistic approach to alignment. This paper describes oblivious agents: agents that are architected in such a way that their effective utility function is an aggregation of a known and hidden sub-functions. The hidden component, to be maximized, is internally implemented as a black box, preventing the agent from examining it. The known component, to be minimized, is knowledge of the hidden sub-function. Architectural constraints further influence how agent actions can evolve its internal environment model. We show that an oblivious agent, behaving rationally, constructs an internal approximation of designers' intentions (i.e., infers alignment), and, as a consequence of its architecture and effective utility function, behaves in such a way that maximizes alignment; i.e., maximizing the approximated intention function. We show that, paradoxically, it does this for whatever utility function is used as the hidden component and, in contrast with extant techniques, chances of alignment actually improve as agent intelligence grows.
{"title":"Agents Need Not Know Their Purpose","authors":"Paulo Garcia","doi":"10.48550/arXiv.2402.09734","DOIUrl":"https://doi.org/10.48550/arXiv.2402.09734","url":null,"abstract":"Ensuring artificial intelligence behaves in such a way that is aligned with human values is commonly referred to as the alignment challenge. Prior work has shown that rational agents, behaving in such a way that maximizes a utility function, will inevitably behave in such a way that is not aligned with human values, especially as their level of intelligence goes up. Prior work has also shown that there is no\"one true utility function\"; solutions must include a more holistic approach to alignment. This paper describes oblivious agents: agents that are architected in such a way that their effective utility function is an aggregation of a known and hidden sub-functions. The hidden component, to be maximized, is internally implemented as a black box, preventing the agent from examining it. The known component, to be minimized, is knowledge of the hidden sub-function. Architectural constraints further influence how agent actions can evolve its internal environment model. We show that an oblivious agent, behaving rationally, constructs an internal approximation of designers' intentions (i.e., infers alignment), and, as a consequence of its architecture and effective utility function, behaves in such a way that maximizes alignment; i.e., maximizing the approximated intention function. We show that, paradoxically, it does this for whatever utility function is used as the hidden component and, in contrast with extant techniques, chances of alignment actually improve as agent intelligence grows.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139962620","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.10191
Xinchi Qiu, Yan Gao, Lorenzo Sani, Heng Pan, Wanru Zhao, Pedro Gusmão, Mina Alibeigi, Alexandru Iacob, Nicholas D. Lane
Federated learning (FL) is a distributed learning paradigm that facilitates collaborative training of a shared global model across devices while keeping data localized. The deployment of FL in numerous real-world applications faces delays, primarily due to the prevalent reliance on supervised tasks. Generating detailed labels at edge devices, if feasible, is demanding, given resource constraints and the imperative for continuous data updates. In addressing these challenges, solutions such as federated semi-supervised learning (FSSL), which relies on unlabeled clients' data and a limited amount of labeled data on the server, become pivotal. In this paper, we propose FedAnchor, an innovative FSSL method that introduces a unique double-head structure, called anchor head, paired with the classification head trained exclusively on labeled anchor data on the server. The anchor head is empowered with a newly designed label contrastive loss based on the cosine similarity metric. Our approach mitigates the confirmation bias and overfitting issues associated with pseudo-labeling techniques based on high-confidence model prediction samples. Extensive experiments on CIFAR10/100 and SVHN datasets demonstrate that our method outperforms the state-of-the-art method by a significant margin in terms of convergence rate and model accuracy.
{"title":"FedAnchor: Enhancing Federated Semi-Supervised Learning with Label Contrastive Loss for Unlabeled Clients","authors":"Xinchi Qiu, Yan Gao, Lorenzo Sani, Heng Pan, Wanru Zhao, Pedro Gusmão, Mina Alibeigi, Alexandru Iacob, Nicholas D. Lane","doi":"10.48550/arXiv.2402.10191","DOIUrl":"https://doi.org/10.48550/arXiv.2402.10191","url":null,"abstract":"Federated learning (FL) is a distributed learning paradigm that facilitates collaborative training of a shared global model across devices while keeping data localized. The deployment of FL in numerous real-world applications faces delays, primarily due to the prevalent reliance on supervised tasks. Generating detailed labels at edge devices, if feasible, is demanding, given resource constraints and the imperative for continuous data updates. In addressing these challenges, solutions such as federated semi-supervised learning (FSSL), which relies on unlabeled clients' data and a limited amount of labeled data on the server, become pivotal. In this paper, we propose FedAnchor, an innovative FSSL method that introduces a unique double-head structure, called anchor head, paired with the classification head trained exclusively on labeled anchor data on the server. The anchor head is empowered with a newly designed label contrastive loss based on the cosine similarity metric. Our approach mitigates the confirmation bias and overfitting issues associated with pseudo-labeling techniques based on high-confidence model prediction samples. Extensive experiments on CIFAR10/100 and SVHN datasets demonstrate that our method outperforms the state-of-the-art method by a significant margin in terms of convergence rate and model accuracy.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139962730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.09797
Hyewon Han, Naveen Kumar
In this work, we propose a novel cross-talk rejection framework for a multi-channel multi-talker setup for a live multiparty interactive show. Our far-field audio setup is required to be hands-free during live interaction and comprises four adjacent talkers with directional microphones in the same space. Such setups often introduce heavy cross-talk between channels, resulting in reduced automatic speech recognition (ASR) and natural language understanding (NLU) performance. To address this problem, we propose voice activity detection (VAD) model for all talkers using multichannel information, which is then used to filter audio for downstream tasks. We adopt a synthetic training data generation approach through playback and re-recording for such scenarios, simulating challenging speech overlap conditions. We train our models on this synthetic data and demonstrate that our approach outperforms single-channel VAD models and energy-based multi-channel VAD algorithm in various acoustic environments. In addition to VAD results, we also present multiparty ASR evaluation results to highlight the impact of using our VAD model for filtering audio in downstream tasks by significantly reducing the insertion error.
在这项工作中,我们为现场多方互动节目的多通道多谈话者设置提出了一种新颖的串扰抑制框架。我们的远场音频设置要求在现场互动时免提,由四个相邻的谈话者在同一空间内使用定向麦克风组成。这种设置通常会在声道之间产生严重的串扰,从而降低自动语音识别(ASR)和自然语言理解(NLU)的性能。为解决这一问题,我们提出了利用多通道信息对所有说话者进行语音活动检测(VAD)的模型,然后利用该模型为下游任务过滤音频。我们采用一种合成训练数据生成方法,通过回放和重新录制此类场景,模拟具有挑战性的语音重叠条件。我们在这些合成数据上训练我们的模型,并证明我们的方法在各种声学环境中优于单通道 VAD 模型和基于能量的多通道 VAD 算法。除了 VAD 结果外,我们还展示了多方 ASR 评估结果,以强调在下游任务中使用我们的 VAD 模型过滤音频的影响,即显著减少插入误差。
{"title":"A cross-talk robust multichannel VAD model for multiparty agent interactions trained using synthetic re-recordings","authors":"Hyewon Han, Naveen Kumar","doi":"10.48550/arXiv.2402.09797","DOIUrl":"https://doi.org/10.48550/arXiv.2402.09797","url":null,"abstract":"In this work, we propose a novel cross-talk rejection framework for a multi-channel multi-talker setup for a live multiparty interactive show. Our far-field audio setup is required to be hands-free during live interaction and comprises four adjacent talkers with directional microphones in the same space. Such setups often introduce heavy cross-talk between channels, resulting in reduced automatic speech recognition (ASR) and natural language understanding (NLU) performance. To address this problem, we propose voice activity detection (VAD) model for all talkers using multichannel information, which is then used to filter audio for downstream tasks. We adopt a synthetic training data generation approach through playback and re-recording for such scenarios, simulating challenging speech overlap conditions. We train our models on this synthetic data and demonstrate that our approach outperforms single-channel VAD models and energy-based multi-channel VAD algorithm in various acoustic environments. In addition to VAD results, we also present multiparty ASR evaluation results to highlight the impact of using our VAD model for filtering audio in downstream tasks by significantly reducing the insertion error.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139962792","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2024-02-15DOI: 10.48550/arXiv.2402.10184
Tianyi Qiu, Fanzhi Zeng, Jiaming Ji, Dong Yan, Kaile Wang, Jiayi Zhou, Han Yang, Josef Dai, Xuehai Pan, Yaodong Yang
There is a trilemma in reinforcement learning from human feedback (RLHF): the incompatibility between highly diverse contexts, low labeling cost, and reliable alignment performance. Here we aim to mitigate such incompatibility through the design of dataset information structures during reward modeling, and meanwhile propose new, generalizable methods of analysis that have wider applications, including potentially shedding light on goal misgeneralization. Specifically, we first reexamine the RLHF process and propose a theoretical framework portraying it as an autoencoding process over text distributions. Our framework formalizes the RLHF objective of ensuring distributional consistency between human preference and large language model (LLM) behavior. Based on this framework, we introduce a new method to model generalization in the reward modeling stage of RLHF, the induced Bayesian network (IBN). Drawing from random graph theory and causal analysis, it enables empirically grounded derivation of generalization error bounds, a key improvement over classical methods of generalization analysis. An insight from our analysis is the superiority of the tree-based information structure in reward modeling, compared to chain-based baselines in conventional RLHF methods. We derive that in complex contexts with limited data, the tree-based reward model (RM) induces up to $Theta(log n/loglog n)$ times less variance than chain-based RM where $n$ is the dataset size. As validation, we demonstrate that on three NLP tasks, the tree-based RM achieves 65% win rate on average against chain-based baselines. Looking ahead, we hope to extend the IBN analysis to help understand the phenomenon of goal misgeneralization.
{"title":"Rethinking Information Structures in RLHF: Reward Generalization from a Graph Theory Perspective","authors":"Tianyi Qiu, Fanzhi Zeng, Jiaming Ji, Dong Yan, Kaile Wang, Jiayi Zhou, Han Yang, Josef Dai, Xuehai Pan, Yaodong Yang","doi":"10.48550/arXiv.2402.10184","DOIUrl":"https://doi.org/10.48550/arXiv.2402.10184","url":null,"abstract":"There is a trilemma in reinforcement learning from human feedback (RLHF): the incompatibility between highly diverse contexts, low labeling cost, and reliable alignment performance. Here we aim to mitigate such incompatibility through the design of dataset information structures during reward modeling, and meanwhile propose new, generalizable methods of analysis that have wider applications, including potentially shedding light on goal misgeneralization. Specifically, we first reexamine the RLHF process and propose a theoretical framework portraying it as an autoencoding process over text distributions. Our framework formalizes the RLHF objective of ensuring distributional consistency between human preference and large language model (LLM) behavior. Based on this framework, we introduce a new method to model generalization in the reward modeling stage of RLHF, the induced Bayesian network (IBN). Drawing from random graph theory and causal analysis, it enables empirically grounded derivation of generalization error bounds, a key improvement over classical methods of generalization analysis. An insight from our analysis is the superiority of the tree-based information structure in reward modeling, compared to chain-based baselines in conventional RLHF methods. We derive that in complex contexts with limited data, the tree-based reward model (RM) induces up to $Theta(log n/loglog n)$ times less variance than chain-based RM where $n$ is the dataset size. As validation, we demonstrate that on three NLP tasks, the tree-based RM achieves 65% win rate on average against chain-based baselines. Looking ahead, we hope to extend the IBN analysis to help understand the phenomenon of goal misgeneralization.","PeriodicalId":8425,"journal":{"name":"ArXiv","volume":null,"pages":null},"PeriodicalIF":0.0,"publicationDate":"2024-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139962930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}