Simultaneous speech translation (SimulST) aims to translate ongoing speech with low latency while preserving translation quality. Recent studies have adopted large language models (LLMs) for SimulST and formulated it as a multi-turn dialogue task for efficient generation, achieving state-of-the-art performance. However...
Synthetic data has become a common resource for training language models, and this trend has spilled over into research on developmentally plausible language modeling. However, it remains unclear whether LM-generated child-adult interactions actually resemble natural ones. We compare the synthetic TinyDialogues dataset...
A bilingual Japanese--English persona--line conflict benchmark in which each item pairs one line with congruent and conflicting personas and pre-declares the expected acoustic direction shows that learned repair of the plan-to-synthesis interface can recover counter-textual delivery, although its effect differs across...
End-to-end full-duplex spoken dialogue models enable natural, real-time spoken interactions with interruptions, backchannels, and overlapping speech. Prior evaluations have assessed diverse interaction capabilities and identified some degradation patterns across conversational turns. However, how model behavior evolves...
Junho Kim, CheHyun Lee, Donghun Kang et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
GLP-1 medications are increasingly used for weight management, and many people share their experiences online. These narratives show why people take the drugs and what they experience, but they remain unstructured and disconnected from formal pharmacovigilance. We therefore present AskGLP1, a system that transforms onl...
Zhaoqian Xue, Anay Somani, Siddhartha Singh et al.· 0 citations
Pretrained transformers spend a large share of their inference compute inside self-attention, yet the per-head query-key interaction is structurally low-rank by construction. We ask a diagnostic question: across layers and architectures, which heads actually use the full rank budget on real data, and which do not? We p...
Debopam Sanyal, Hongjie Chen, Riya Elizabeth Kuriyan et al.· 0 citations
Outcome RL raises pass@1 but not pass@k at large k (Yue et al., 2025). But every pass@1 is scored under a token budget. Every recipe also trains under one. On physics and math problems at fixed token budgets, we compare the pass@1 of 20 released RL models (1.5B to 32B) with that of the thinking models they started from...
Controlling personality in large language models is important for building conversational agents that can be customized to different users and interaction settings. However, existing personality control methods often face a tradeoff between strong trait expression and natural dialogue generation. We introduce FATE-5, F...
Gunhee Cho, Hongyeob Kim, Yun-Gyung Cheong· 0 citations
A numerical claim generated by a financial analyst agent can match a number in a company filing but refer to the wrong entity, concept, period, unit or scope. A check based on reference answers can mark such an answer as wrong but does not explain why it fails. In this paper, we present NumTrace, a new benchmark that a...
Tian-Lei Zhu, Matt White, Xue Liu et al.· 0 citations
Rubric-based evaluation asks which level of a rubric a response meets. It is usually done by prompting a large language model (LLM) as a judge, which needs a generative call per judgment and reads every level description, so its cost grows with the rubric. We present RubriX, a rubric judge whose architecture is inspire...
Retrieval-augmented generation (RAG) can improve factual grounding without guaranteeing that every generated claim is supported. Previous work has studied claim-level conformal factuality in single-stage RAG and, more recently, multi-hop RAG. We examine its behavior in agentic RAG, where a language-model policy adaptiv...
Much of the information needed for data analysis is buried in text columns: an analyst may infer customer frustration from complaint text and use it to explain resolution time. We call workloads in which processing text columns is central to subsequent data analysis TADA (\emph{Text Analysis in Data Analytics}). In an...
Jin-Xiang Xie, Zi-Hao Li, Rui Ding et al.· 0 citations
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.
