Curating the frontier of Artificial Intelligence

Multi-LCB: Extending LiveCodeBench to Multiple Programming Languages

LiveCodeBench (LCB) 最近已成为广泛采用的基准，用于在代码生成任务上评估大型语言模型 (LLMs)。通过策划...

LiveCodeBench (LCB) has recently become a widely adopted benchmark for evaluating large language models (LLMs) on code-generation tasks. By curating c...

FlowEdit: Associative Memory for Lifelong Pronunciation Adaptation in Flow-Matching TTS

流匹配文本到语音系统实现了卓越的零样本质量，但在部署后保持静态: 词汇外发音错误...

Flow-matching text-to-speech systems achieve remarkable zero-shot quality but remain static after deployment: pronunciation errors on out-of-vocabular...

DeepSWIP: Quotient-WMC Counterfactuals for Neural Probabilistic Logic Programs

DeepProbLog 等神经符号系统将神经感知与概率逻辑相结合,，但标准推理是关联的。反事实...

Neurosymbolic systems such as DeepProbLog combine neural perception with probabilistic logic, but standard inference is associational. Counterfactual ...

LedgerAgent: Structured State for Policy-Adherent Tool-Calling Agents

客户服务域中遵守策略的工具调用代理必须在调用工具并遵守域策略的同时保持轮流任务状态......

Policy-adherent tool-calling agents in customer-service domains must maintain task states across turns while calling tools and obeying domain policies...

How Do Instructions Shape Speech? Cross-Attention Attribution for Style-Captioned Text-to-Speech

样式字幕文本转语音系统使用自然语言来控制语音特征,，但单个单词如何影响声音输出仍然是......

Style-captioned text-to-speech systems use natural language to control voice characteristics, but how individual words influence acoustic output remai...

Toward Calibrated Mixture-of-Experts Under Distribution Shift

校准将模型的的预测不确定性与其经验结果的频率保持一致，对于理解和信任代表很重要......

Calibration aligns a model's predictive uncertainty with the frequencies of its empirical outcomes and is important for understanding and trusting rep...

A better way to model the behavior of metal alloys

从事航空航天,、能源, 和计算领域前沿工作的公司不断寻找新材料来提高性能。但为了...

Companies working at the frontier of aerospace, energy, and computing are constantly looking for new materials to improve performance. But in order to...

TxBench-PP: Analyzing AI Agent Performance on Small-Molecule Preclinical Pharmacology

人工智能(AI)代理承诺通过压缩解释和决策循环来加速药物发现,但实际部署...

Artificial intelligence (AI) agents promise to accelerate drug discovery by compressing interpretation and decision-making loops, but practical deploy...

X+Slides: Benchmarking Audience-Conditioned Slide Generation

从源文档自动生成幻灯片是大型语言模型(LLMs) 的一个重要应用。现有基准主要...

Automatically generating slide decks from source documents is an important application of large language models (LLMs). Existing benchmarks primarily ...

NeSyCat Torch: A Differentiable Tensor Implementation of Categorical Semantics for Neurosymbolic Learning

神经符号语义是支离破碎的:经典,模糊,概率和神经系统每个都通过自己的归纳规则定义真理。 NeSyCat, 扩展...

Neurosymbolic semantics is fragmented: classical, fuzzy, probabilistic and neural systems each define truth by their own inductive rules. NeSyCat, ext...