VoiceAgentBench: Are Voice Assistants ready for agentic tasks?cs.AI updates on arXiv.org arXiv:2510.07978v2 Announce Type: replace
Abstract: Large-scale Speech Language Models (SpeechLMs) have enabled voice assistants capable of understanding natural spoken queries and performing complex tasks. However, existing speech benchmarks primarily focus on isolated capabilities such as transcription, or question-answering, and do not systematically evaluate agentic scenarios encompassing multilingual and cultural understanding, as well as adversarial robustness. To address this, we introduce VoiceAgentBench, a comprehensive benchmark designed to evaluate SpeechLMs in realistic spoken agentic settings. It comprises over 5,500 synthetic spoken queries, including dialogues grounded in Indian context, covering single-tool invocations, multi-tool workflows, multi-turn interactions, and safety evaluations. The benchmark supports English, Hindi, and 5 other Indian languages, reflecting real-world linguistic and cultural diversity. We simulate speaker variability using a novel sampling algorithm that selects audios for TTS voice conversion based on its speaker embeddings, maximizing acoustic and speaker diversity. Our evaluation measures tool selection accuracy, structural consistency, and the correctness of tool invocations, including adversarial robustness. Our experiments reveal significant gaps in contextual tool orchestration tasks, Indic generalization, and adversarial robustness, exposing critical limitations of current SpeechLMs.
arXiv:2510.07978v2 Announce Type: replace
Abstract: Large-scale Speech Language Models (SpeechLMs) have enabled voice assistants capable of understanding natural spoken queries and performing complex tasks. However, existing speech benchmarks primarily focus on isolated capabilities such as transcription, or question-answering, and do not systematically evaluate agentic scenarios encompassing multilingual and cultural understanding, as well as adversarial robustness. To address this, we introduce VoiceAgentBench, a comprehensive benchmark designed to evaluate SpeechLMs in realistic spoken agentic settings. It comprises over 5,500 synthetic spoken queries, including dialogues grounded in Indian context, covering single-tool invocations, multi-tool workflows, multi-turn interactions, and safety evaluations. The benchmark supports English, Hindi, and 5 other Indian languages, reflecting real-world linguistic and cultural diversity. We simulate speaker variability using a novel sampling algorithm that selects audios for TTS voice conversion based on its speaker embeddings, maximizing acoustic and speaker diversity. Our evaluation measures tool selection accuracy, structural consistency, and the correctness of tool invocations, including adversarial robustness. Our experiments reveal significant gaps in contextual tool orchestration tasks, Indic generalization, and adversarial robustness, exposing critical limitations of current SpeechLMs. Read More
FaStfact: Faster, Stronger Long-Form Factuality Evaluations in LLMscs.AI updates on arXiv.org arXiv:2510.12839v2 Announce Type: replace-cross
Abstract: Evaluating the factuality of long-form generations from Large Language Models (LLMs) remains challenging due to efficiency bottlenecks and reliability concerns. Prior efforts attempt this by decomposing text into claims, searching for evidence, and verifying claims, but suffer from critical drawbacks: (1) inefficiency due to overcomplicated pipeline components, and (2) ineffectiveness stemming from inaccurate claim sets and insufficient evidence. To address these limitations, we propose textbf{FaStfact}, an evaluation framework that achieves the highest alignment with human evaluation and time/token efficiency among existing baselines. FaStfact first employs chunk-level claim extraction integrated with confidence-based pre-verification, significantly reducing the time and token cost while ensuring reliability. For searching and verification, it collects document-level evidence from crawled web-pages and selectively retrieves it during verification. Extensive experiments based on an annotated benchmark textbf{FaStfact-Bench} demonstrate the reliability of FaStfact in both efficiently and effectively evaluating long-form factuality. Code, benchmark data, and annotation interface tool are available at https://github.com/Yingjia-Wan/FaStfact.
arXiv:2510.12839v2 Announce Type: replace-cross
Abstract: Evaluating the factuality of long-form generations from Large Language Models (LLMs) remains challenging due to efficiency bottlenecks and reliability concerns. Prior efforts attempt this by decomposing text into claims, searching for evidence, and verifying claims, but suffer from critical drawbacks: (1) inefficiency due to overcomplicated pipeline components, and (2) ineffectiveness stemming from inaccurate claim sets and insufficient evidence. To address these limitations, we propose textbf{FaStfact}, an evaluation framework that achieves the highest alignment with human evaluation and time/token efficiency among existing baselines. FaStfact first employs chunk-level claim extraction integrated with confidence-based pre-verification, significantly reducing the time and token cost while ensuring reliability. For searching and verification, it collects document-level evidence from crawled web-pages and selectively retrieves it during verification. Extensive experiments based on an annotated benchmark textbf{FaStfact-Bench} demonstrate the reliability of FaStfact in both efficiently and effectively evaluating long-form factuality. Code, benchmark data, and annotation interface tool are available at https://github.com/Yingjia-Wan/FaStfact. Read More
AI for Requirements Engineering: Industry adoption and Practitioner perspectivescs.AI updates on arXiv.org arXiv:2511.01324v3 Announce Type: replace-cross
Abstract: The integration of AI for Requirements Engineering (RE) presents significant benefits but also poses real challenges. Although RE is fundamental to software engineering, limited research has examined AI adoption in RE. We surveyed 55 software practitioners to map AI usage across four RE phases: Elicitation, Analysis, Specification, and Validation, and four approaches for decision making: human-only decisions, AI validation, Human AI Collaboration (HAIC), and full AI automation. Participants also shared their perceptions, challenges, and opportunities when applying AI for RE tasks. Our data show that 58.2% of respondents already use AI in RE, and 69.1% view its impact as positive or very positive. HAIC dominates practice, accounting for 54.4% of all RE techniques, while full AI automation remains minimal at 5.4%. Passive AI validation (4.4 to 6.2%) lags even further behind, indicating that practitioners value AI’s active support over passive oversight. These findings suggest that AI is most effective when positioned as a collaborative partner rather than a replacement for human expertise. It also highlights the need for RE-specific HAIC frameworks along with robust and responsible AI governance as AI adoption in RE grows.
arXiv:2511.01324v3 Announce Type: replace-cross
Abstract: The integration of AI for Requirements Engineering (RE) presents significant benefits but also poses real challenges. Although RE is fundamental to software engineering, limited research has examined AI adoption in RE. We surveyed 55 software practitioners to map AI usage across four RE phases: Elicitation, Analysis, Specification, and Validation, and four approaches for decision making: human-only decisions, AI validation, Human AI Collaboration (HAIC), and full AI automation. Participants also shared their perceptions, challenges, and opportunities when applying AI for RE tasks. Our data show that 58.2% of respondents already use AI in RE, and 69.1% view its impact as positive or very positive. HAIC dominates practice, accounting for 54.4% of all RE techniques, while full AI automation remains minimal at 5.4%. Passive AI validation (4.4 to 6.2%) lags even further behind, indicating that practitioners value AI’s active support over passive oversight. These findings suggest that AI is most effective when positioned as a collaborative partner rather than a replacement for human expertise. It also highlights the need for RE-specific HAIC frameworks along with robust and responsible AI governance as AI adoption in RE grows. Read More
Balancing Tails when Comparing Distributions: Comprehensive Equity Index (CEI) with Application to Bias Evaluation in Operational Face Biometricscs.AI updates on arXiv.org arXiv:2506.10564v2 Announce Type: replace-cross
Abstract: Demographic bias in high-performance face recognition (FR) systems often eludes detection by existing metrics, especially with respect to subtle disparities in the tails of the score distribution. We introduce the Comprehensive Equity Index (CEI), a novel metric designed to address this limitation. CEI uniquely analyzes genuine and impostor score distributions separately, enabling a configurable focus on tail probabilities while also considering overall distribution shapes. Our extensive experiments (evaluating state-of-the-art FR systems, intentionally biased models, and diverse datasets) confirm CEI’s superior ability to detect nuanced biases where previous methods fall short. Furthermore, we present CEI^A, an automated version of the metric that enhances objectivity and simplifies practical application. CEI provides a robust and sensitive tool for operational FR fairness assessment. The proposed methods have been developed particularly for bias evaluation in face biometrics but, in general, they are applicable for comparing statistical distributions in any problem where one is interested in analyzing the distribution tails.
arXiv:2506.10564v2 Announce Type: replace-cross
Abstract: Demographic bias in high-performance face recognition (FR) systems often eludes detection by existing metrics, especially with respect to subtle disparities in the tails of the score distribution. We introduce the Comprehensive Equity Index (CEI), a novel metric designed to address this limitation. CEI uniquely analyzes genuine and impostor score distributions separately, enabling a configurable focus on tail probabilities while also considering overall distribution shapes. Our extensive experiments (evaluating state-of-the-art FR systems, intentionally biased models, and diverse datasets) confirm CEI’s superior ability to detect nuanced biases where previous methods fall short. Furthermore, we present CEI^A, an automated version of the metric that enhances objectivity and simplifies practical application. CEI provides a robust and sensitive tool for operational FR fairness assessment. The proposed methods have been developed particularly for bias evaluation in face biometrics but, in general, they are applicable for comparing statistical distributions in any problem where one is interested in analyzing the distribution tails. Read More
REFA: Reference Free Alignment for multi-preference optimizationcs.AI updates on arXiv.org arXiv:2412.16378v4 Announce Type: replace-cross
Abstract: To mitigate reward hacking from response verbosity, modern preference optimization methods are increasingly adopting length normalization (e.g., SimPO, ORPO, LN-DPO). While effective against this bias, we demonstrate that length normalization itself introduces a failure mode: the URSLA shortcut. Here models learn to satisfy the alignment objective by prematurely truncating low-quality responses rather than learning from their semantic content. To address this, we introduce REFA, a new alignment framework that proposes probabilistic control on a structural token that controls termination. Our core innovation is a new class of regularizers that operate directly on the probability of the End-of-Sequence (EOS) token, a previously unexploited control lever. This token-level intervention provides a principled solution to the URSLA shortcut, ensuring genuine quality improvements. Furthermore, it unlocks a versatile mechanism for managing the alignment-efficiency tradeoff, enabling practitioners to fine-tune models that adhere to specific token budgets. Empirically, REFA achieves a 60.29% win rate and a 52.17% length-controlled win rate on AlpacaEval2 with Llama-3-8B-Instruct, demonstrating the power of our token-level control paradigm.
arXiv:2412.16378v4 Announce Type: replace-cross
Abstract: To mitigate reward hacking from response verbosity, modern preference optimization methods are increasingly adopting length normalization (e.g., SimPO, ORPO, LN-DPO). While effective against this bias, we demonstrate that length normalization itself introduces a failure mode: the URSLA shortcut. Here models learn to satisfy the alignment objective by prematurely truncating low-quality responses rather than learning from their semantic content. To address this, we introduce REFA, a new alignment framework that proposes probabilistic control on a structural token that controls termination. Our core innovation is a new class of regularizers that operate directly on the probability of the End-of-Sequence (EOS) token, a previously unexploited control lever. This token-level intervention provides a principled solution to the URSLA shortcut, ensuring genuine quality improvements. Furthermore, it unlocks a versatile mechanism for managing the alignment-efficiency tradeoff, enabling practitioners to fine-tune models that adhere to specific token budgets. Empirically, REFA achieves a 60.29% win rate and a 52.17% length-controlled win rate on AlpacaEval2 with Llama-3-8B-Instruct, demonstrating the power of our token-level control paradigm. Read More
PlanU: Large Language Model Reasoning through Planning under Uncertaintycs.AI updates on arXiv.org arXiv:2510.18442v2 Announce Type: replace
Abstract: Large Language Models (LLMs) are increasingly being explored across a range of reasoning tasks. However, LLMs sometimes struggle with reasoning tasks under uncertainty that are relatively easy for humans, such as planning actions in stochastic environments. The adoption of LLMs for reasoning is impeded by uncertainty challenges, such as LLM uncertainty and environmental uncertainty. LLM uncertainty arises from the stochastic sampling process inherent to LLMs. Most LLM-based Decision-Making (LDM) approaches address LLM uncertainty through multiple reasoning chains or search trees. However, these approaches overlook environmental uncertainty, which leads to poor performance in environments with stochastic state transitions. Some recent LDM approaches deal with uncertainty by forecasting the probability of unknown variables. However, they are not designed for multi-step reasoning tasks that require interaction with the environment. To address uncertainty in LLM decision-making, we introduce PlanU, an LLM-based planning method that captures uncertainty within Monte Carlo Tree Search (MCTS). PlanU models the return of each node in the MCTS as a quantile distribution, which uses a set of quantiles to represent the return distribution. To balance exploration and exploitation during tree search, PlanU introduces an Upper Confidence Bounds with Curiosity (UCC) score which estimates the uncertainty of MCTS nodes. Through extensive experiments, we demonstrate the effectiveness of PlanU in LLM-based reasoning tasks under uncertainty.
arXiv:2510.18442v2 Announce Type: replace
Abstract: Large Language Models (LLMs) are increasingly being explored across a range of reasoning tasks. However, LLMs sometimes struggle with reasoning tasks under uncertainty that are relatively easy for humans, such as planning actions in stochastic environments. The adoption of LLMs for reasoning is impeded by uncertainty challenges, such as LLM uncertainty and environmental uncertainty. LLM uncertainty arises from the stochastic sampling process inherent to LLMs. Most LLM-based Decision-Making (LDM) approaches address LLM uncertainty through multiple reasoning chains or search trees. However, these approaches overlook environmental uncertainty, which leads to poor performance in environments with stochastic state transitions. Some recent LDM approaches deal with uncertainty by forecasting the probability of unknown variables. However, they are not designed for multi-step reasoning tasks that require interaction with the environment. To address uncertainty in LLM decision-making, we introduce PlanU, an LLM-based planning method that captures uncertainty within Monte Carlo Tree Search (MCTS). PlanU models the return of each node in the MCTS as a quantile distribution, which uses a set of quantiles to represent the return distribution. To balance exploration and exploitation during tree search, PlanU introduces an Upper Confidence Bounds with Curiosity (UCC) score which estimates the uncertainty of MCTS nodes. Through extensive experiments, we demonstrate the effectiveness of PlanU in LLM-based reasoning tasks under uncertainty. Read More
Learning Under Laws: A Constraint-Projected Neural PDE Solver that Eliminates Hallucinationscs.AI updates on arXiv.org arXiv:2511.03578v1 Announce Type: cross
Abstract: Neural networks can approximate solutions to partial differential equations, but they often break the very laws they are meant to model-creating mass from nowhere, drifting shocks, or violating conservation and entropy. We address this by training within the laws of physics rather than beside them. Our framework, called Constraint-Projected Learning (CPL), keeps every update physically admissible by projecting network outputs onto the intersection of constraint sets defined by conservation, Rankine-Hugoniot balance, entropy, and positivity. The projection is differentiable and adds only about 10% computational overhead, making it fully compatible with back-propagation. We further stabilize training with total-variation damping (TVD) to suppress small oscillations and a rollout curriculum that enforces consistency over long prediction horizons. Together, these mechanisms eliminate both hard and soft violations: conservation holds at machine precision, total-variation growth vanishes, and entropy and error remain bounded. On Burgers and Euler systems, CPL produces stable, physically lawful solutions without loss of accuracy. Instead of hoping neural solvers will respect physics, CPL makes that behavior an intrinsic property of the learning process.
arXiv:2511.03578v1 Announce Type: cross
Abstract: Neural networks can approximate solutions to partial differential equations, but they often break the very laws they are meant to model-creating mass from nowhere, drifting shocks, or violating conservation and entropy. We address this by training within the laws of physics rather than beside them. Our framework, called Constraint-Projected Learning (CPL), keeps every update physically admissible by projecting network outputs onto the intersection of constraint sets defined by conservation, Rankine-Hugoniot balance, entropy, and positivity. The projection is differentiable and adds only about 10% computational overhead, making it fully compatible with back-propagation. We further stabilize training with total-variation damping (TVD) to suppress small oscillations and a rollout curriculum that enforces consistency over long prediction horizons. Together, these mechanisms eliminate both hard and soft violations: conservation holds at machine precision, total-variation growth vanishes, and entropy and error remain bounded. On Burgers and Euler systems, CPL produces stable, physically lawful solutions without loss of accuracy. Instead of hoping neural solvers will respect physics, CPL makes that behavior an intrinsic property of the learning process. Read More
Comparing the Performance of LLMs in RAG-based Question-Answering: A Case Study in Computer Science Literaturecs.AI updates on arXiv.org arXiv:2511.03261v1 Announce Type: cross
Abstract: Retrieval Augmented Generation (RAG) is emerging as a powerful technique to enhance the capabilities of Generative AI models by reducing hallucination. Thus, the increasing prominence of RAG alongside Large Language Models (LLMs) has sparked interest in comparing the performance of different LLMs in question-answering (QA) in diverse domains. This study compares the performance of four open-source LLMs, Mistral-7b-instruct, LLaMa2-7b-chat, Falcon-7b-instruct and Orca-mini-v3-7b, and OpenAI’s trending GPT-3.5 over QA tasks within the computer science literature leveraging RAG support. Evaluation metrics employed in the study include accuracy and precision for binary questions and ranking by a human expert, ranking by Google’s AI model Gemini, alongside cosine similarity for long-answer questions. GPT-3.5, when paired with RAG, effectively answers binary and long-answer questions, reaffirming its status as an advanced LLM. Regarding open-source LLMs, Mistral AI’s Mistral-7b-instruct paired with RAG surpasses the rest in answering both binary and long-answer questions. However, among the open-source LLMs, Orca-mini-v3-7b reports the shortest average latency in generating responses, whereas LLaMa2-7b-chat by Meta reports the highest average latency. This research underscores the fact that open-source LLMs, too, can go hand in hand with proprietary models like GPT-3.5 with better infrastructure.
arXiv:2511.03261v1 Announce Type: cross
Abstract: Retrieval Augmented Generation (RAG) is emerging as a powerful technique to enhance the capabilities of Generative AI models by reducing hallucination. Thus, the increasing prominence of RAG alongside Large Language Models (LLMs) has sparked interest in comparing the performance of different LLMs in question-answering (QA) in diverse domains. This study compares the performance of four open-source LLMs, Mistral-7b-instruct, LLaMa2-7b-chat, Falcon-7b-instruct and Orca-mini-v3-7b, and OpenAI’s trending GPT-3.5 over QA tasks within the computer science literature leveraging RAG support. Evaluation metrics employed in the study include accuracy and precision for binary questions and ranking by a human expert, ranking by Google’s AI model Gemini, alongside cosine similarity for long-answer questions. GPT-3.5, when paired with RAG, effectively answers binary and long-answer questions, reaffirming its status as an advanced LLM. Regarding open-source LLMs, Mistral AI’s Mistral-7b-instruct paired with RAG surpasses the rest in answering both binary and long-answer questions. However, among the open-source LLMs, Orca-mini-v3-7b reports the shortest average latency in generating responses, whereas LLaMa2-7b-chat by Meta reports the highest average latency. This research underscores the fact that open-source LLMs, too, can go hand in hand with proprietary models like GPT-3.5 with better infrastructure. Read More
Decoupling Augmentation Bias in Prompt Learning for Vision-Language Modelscs.AI updates on arXiv.org arXiv:2511.03367v1 Announce Type: cross
Abstract: Recent advances in large-scale vision and language models have led to significant progress in zero-shot learning tasks. Methods such as CoOp and CoCoOp have shown that replacing handcrafted prompts with learnable vectors, known as prompt learning, can result in improved performance. However, these models often struggle to generalize to entirely unseen categories. While traditional zero-shot learning techniques benefit from various data augmentation strategies, prompt learning has primarily focused on text-based modifications, leaving the potential of image-based augmentation largely unexplored. In this work, we explore how image-level augmentations, particularly those that introduce attribute-specific variations, can support and enhance prompt learning. Our analysis examines the interaction between these augmentations and soft prompt frameworks, revealing their potential to improve generalization. We also identify a limitation in existing methods, such as CoCoOp, which do not provide explicit guidance for learning prompts that focus on semantically meaningful visual features. To address this, we propose Adding Attributes to Prompt Learning, AAPL, a novel method that introduces adversarial token embeddings to decouple superficial visual variations introduced by augmentation from class-relevant semantic representations. This decoupling enables the learned prompts to concentrate on visually discriminative features that align with the target categories. We conduct comprehensive experiments on eleven benchmark datasets, and AAPL consistently outperforms existing methods across few-shot, zero-shot, cross-dataset, and domain generalization settings. Our source code is publicly available at: https://github.com/Gahyeonkim09/AAPL
arXiv:2511.03367v1 Announce Type: cross
Abstract: Recent advances in large-scale vision and language models have led to significant progress in zero-shot learning tasks. Methods such as CoOp and CoCoOp have shown that replacing handcrafted prompts with learnable vectors, known as prompt learning, can result in improved performance. However, these models often struggle to generalize to entirely unseen categories. While traditional zero-shot learning techniques benefit from various data augmentation strategies, prompt learning has primarily focused on text-based modifications, leaving the potential of image-based augmentation largely unexplored. In this work, we explore how image-level augmentations, particularly those that introduce attribute-specific variations, can support and enhance prompt learning. Our analysis examines the interaction between these augmentations and soft prompt frameworks, revealing their potential to improve generalization. We also identify a limitation in existing methods, such as CoCoOp, which do not provide explicit guidance for learning prompts that focus on semantically meaningful visual features. To address this, we propose Adding Attributes to Prompt Learning, AAPL, a novel method that introduces adversarial token embeddings to decouple superficial visual variations introduced by augmentation from class-relevant semantic representations. This decoupling enables the learned prompts to concentrate on visually discriminative features that align with the target categories. We conduct comprehensive experiments on eleven benchmark datasets, and AAPL consistently outperforms existing methods across few-shot, zero-shot, cross-dataset, and domain generalization settings. Our source code is publicly available at: https://github.com/Gahyeonkim09/AAPL Read More
DQN Performance with Epsilon Greedy Policies and Prioritized Experience Replaycs.AI updates on arXiv.org arXiv:2511.03670v1 Announce Type: cross
Abstract: We present a detailed study of Deep Q-Networks in finite environments, emphasizing the impact of epsilon-greedy exploration schedules and prioritized experience replay. Through systematic experimentation, we evaluate how variations in epsilon decay schedules affect learning efficiency, convergence behavior, and reward optimization. We investigate how prioritized experience replay leads to faster convergence and higher returns and show empirical results comparing uniform, no replay, and prioritized strategies across multiple simulations. Our findings illuminate the trade-offs and interactions between exploration strategies and memory management in DQN training, offering practical recommendations for robust reinforcement learning in resource-constrained settings.
arXiv:2511.03670v1 Announce Type: cross
Abstract: We present a detailed study of Deep Q-Networks in finite environments, emphasizing the impact of epsilon-greedy exploration schedules and prioritized experience replay. Through systematic experimentation, we evaluate how variations in epsilon decay schedules affect learning efficiency, convergence behavior, and reward optimization. We investigate how prioritized experience replay leads to faster convergence and higher returns and show empirical results comparing uniform, no replay, and prioritized strategies across multiple simulations. Our findings illuminate the trade-offs and interactions between exploration strategies and memory management in DQN training, offering practical recommendations for robust reinforcement learning in resource-constrained settings. Read More
