MemoryArena is introduced, a unified evaluation gym for benchmarking agent memory in multi-session Memory-Agent-Environment loops with explicitly interdependent subtasks, where agents must learn from earlier actions and feedback by distilling experiences into memory, and subsequently use that memory to guide later acti...
This paper reformulates MABSA as a generative and explainable task, proposing a unified framework that simultaneously predicts aspect-level sentiment and generates natural language explanations and proposes a dependency-syntax-guided sentiment cue strategy.
Zhongzheng Wang, Yuanhe Tian, Hongzhi Wang et al.· arXiv.org· 2 citations
This paper introduces SoCE, a principled approach for model souping to overcome this shortcoming in precision-sensitive LLM capabilities and demonstrates that SoCE improves performance and robustness across multiple domains and achieves state-of-the-art results on the Berkeley Function Calling Leaderboard.
Shalini Maiti, A. Budhiraja, Bhavul Gauri et al.· arXiv.org· 4 citations
Large Language Models (LLMs) have achieved remarkable success but face significant computational and memory challenges, particularly due to their extensive output vocabularies. The final linear projection layer, mapping hidden states to vocabulary-sized logits, often constitutes a substantial portion of the model's par...
Jintian Shao, Hongyi Huang, Jiayi Wu et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
This work proposes a deployment-friendly approach that encodes cultural behavior in short, tunable soft prompts optimized with Differential Evolution (DE), while keeping model weights frozen and requiring no preference data.
Reem I. Masoud, Martin Ferianc, Philip C. Treleaven et al.· arXiv.org· 4 citations· ⚡1
A bi-modal speech-level transformer is proposed to avoid segment-level labelling and a hierarchical interpretation approach is introduced to provide both speech- level and sentence-level interpretations, based on gradient-weighted attention maps derived from all attention layers to track interactions between input feat...
Qing Deng, S. Luz, Sofia de la Fuente Garcia· arXiv.org· 1 citation
Defenses against jailbreak attacks on Large Language Models (LLMs) operate at different pipeline stages, such as input modification or output guard, but it remains unclear which defenses to deploy at each stage and how to combine them. Prior empirical studies, fragmented by inconsistent attack-success-rate definitions...
We present the Spoken Wikipedia Presentation Corpus, an extension of the Spoken Wikipedia Corpora featuring LLM-generated slide decks for multimodal ASR. Slides are created from LLM-segmented sections using a hybrid pipeline that combines LLM-based content planning with rule-based design decisions. For each section, an...
Thomas Ranzenberger, Steffen Freisinger, Tobias Bocklet et al.· 0 citations
Conformal Privacy Auditing is introduced, a distribution-free calibration framework that provides a statistical certificate of re-identification risk for each released document against LLM-empowered adversaries and enables audits of open-source models and proprietary API models in a unified framework.
Shuo Huang, G. Haffari, Xing-Liang Yuan et al.· Proceedings of the 2026 Conf...· 0 citations
Interrupt and Silent Modeling (ISM), a model-agnostic paradigm that embeds proactive decisions into LLM decoding via two special tokens, is proposed, capturing four states: onset detection, sustained-relevance triggering, irrelevance suppression, and de-duplication.
Amit Kumar Singh Yadav, Ritvik Shrivastava, Xuan Zhang et al.· 1 citation
This work performs experiments on VoxPopuli and The Polish Parliamentary speech datasets and estimates word error rate (WER) differences for different text normalization approaches, showing that the difference due to the lack of number normalization in WER may be substantial.
Voice-Light is presented, a full-duplex cascaded voice agent that combines immediate acoustic onset, a causal adapter sharing a streaming ASR encoder, reversible playback control, and private speculative response generation.
Flint-interactive gives AI agents a language for creating interactive visualizations. Click a chart, filter data, highlight related content, and explore beyond the view the author originally chose. The post Flint-interactive: A language for creating interactive data visualizations appeared first on Microsoft Research.
A new computational approach reveals how cells communicate to assemble into tissues. This could accelerate stem cell engineering, improve disease models, and inform therapies.
Microsoft Research Blog· microsoft.comOct 15, 2026
Zero-knowledge proofs can verify computing results without rerunning workloads or exposing sensitive data. Magnetar helps bring them into the mainstream with support for standard programming languages, major speed gains, and quantum-resistant security. The post Proof-based general-purpose computing with Magnetar appeared first on Microsoft Research.
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
