AI 前沿

聚合全球 AI 前沿资讯与洞察,覆盖大模型、机器学习、自然语言处理等最新动态。
今日 24 小时时间轴 →
排序
标签
大模型1,931机器学习980人工智能817强化学习613AI467AI安全435语言模型413大语言模型401自然语言处理386OpenAI370
ReflectRec: Cycle-Consistency Preference Verification for LLM-based Recommendation

ReflectRec: Cycle-Consistency Preference Verification for LLM-based Recommendation

arXiv自然语言 2026-11-19 13:00 原文 0 阅读 热度 60
CoT-style LLM recommenders generate a natural-language preference note as a chain-of-thought prefix to ranking, but the note is trusted without verification, and the ranker never consults behaviorally...
大模型推荐系统思维链
Coverage Governs Accuracy: When to Expand the Lexicon and When to Fall Back in Low-Resource Manchu OCR

Coverage Governs Accuracy: When to Expand the Lexicon and When to Fall Back in Low-Resource Manchu OCR

arXiv自然语言 2026-11-19 13:00 原文 0 阅读 热度 35
Character-level OCR for low-resource Manchu often produces near-miss spelling errors. Lexicon-constrained decoding can correct such errors with lexical knowledge, but a fixed dictionary cannot cover t...
文字识别满语低资源
NarrowSpec: Draft Hidden Width Need Not Match the Target in Speculative Decoding

NarrowSpec: Draft Hidden Width Need Not Match the Target in Speculative Decoding

arXiv自然语言 2026-11-19 13:00 原文 0 阅读 热度 55
Speculative decoding is widely used to speed up generation in large language models (LLMs), with a small draft model proposing tokens that one forward pass of the target model verifies. Draft models t...
推测解码大语言模型推理加速
Beyond Replacement: Levenshtein Editing for Diffusion Language Models

Beyond Replacement: Levenshtein Editing for Diffusion Language Models

arXiv自然语言 2026-11-19 13:00 原文 0 阅读 热度 55
Diffusion language models (DLMs) offer a flexible alternative to autoregressive models, enabling any-order infilling. While most existing decoding methods treat decoded tokens as fixed, some recent mo...
扩散语言模型文本编辑解码技术
FutureBench: A Renewable Framework for Evaluating Knowledge-Context Conflict in LLMs

FutureBench: A Renewable Framework for Evaluating Knowledge-Context Conflict in LLMs

arXiv自然语言 2026-11-19 13:00 原文 0 阅读 热度 65
Asking a model to check retrieved text can also tell it which source to trust. We introduce FutureBench, a renewable framework for evaluating knowledge-context conflict using timestamped questions, co...
大模型评测框架知识冲突
Can LLMs Understand What They Predict? Evaluating Latent Causal Structure Inference in Financial Markets

Can LLMs Understand What They Predict? Evaluating Latent Causal Structure Inference in Financial Markets

arXiv自然语言 2026-11-19 13:00 原文 1 阅读 热度 75
Large language models (LLMs) are increasingly evaluated through their ability to predict future outcomes, implicitly assuming that predictive success reflects an understanding of the underlying system...
大模型金融市场因果推断
From Arraignment to Resolution: Transformer Models of Court Cases

From Arraignment to Resolution: Transformer Models of Court Cases

arXiv自然语言 2026-11-19 13:00 原文 3 阅读 热度 72
The criminal legal system is composed of many complex decision points, such as which charges are pursued, how bond is determined, plea deal negotiations, and sentencing. Given the profound impact of s...
Transformer模型法律AI司法应用
Tracking Representations in LLM Reasoning Traces for Iterated Games

Tracking Representations in LLM Reasoning Traces for Iterated Games

arXiv自然语言 2026-11-19 13:00 原文 3 阅读 热度 45
Large language model (LLM) agents achieve state-of-the-art performance across domains through chain-of-thought (CoT) reasoning. Despite their ubiquitous use, little is understood about the relationshi...
大语言模型思维链智能体
Chorus: Shared N-gram Embeddings with Layer-wise Adaptation

Chorus: Shared N-gram Embeddings with Layer-wise Adaptation

arXiv自然语言 2026-11-19 13:00 原文 3 阅读 热度 68
Language models can gain capacity at little computational cost from n-gram embedding tables: large tables of learned vectors for short sequences of consecutive tokens, which are looked up and added to...
语言模型词嵌入模型优化
Prioritize, Don't Decide: What We Learned Auditing a Deployed Human-in-the-Loop CV Sorter

Prioritize, Don't Decide: What We Learned Auditing a Deployed Human-in-the-Loop CV Sorter

arXiv自然语言 2026-11-19 13:00 原文 5 阅读 热度 65
A single academic vacancy can draw hundreds of applications, and the first screening round consumes most of a committee's time. We describe a CV Priority Sorter deployed at a national platform for aca...
计算机视觉人机协同简历筛选
AI Will Not Make You Rich (2025)

AI Will Not Make You Rich (2025)

Hacker News 2026-11-19 09:51 原文 17 阅读 热度 45
The disruption is real. It's also predictable.
人工智能行业预测商业分析
Elon Musk tries to explain what’s going on with Ukraine and Starlink

Elon Musk tries to explain what’s going on with Ukraine and Starlink

The Verge 2026-11-19 09:39 原文 16 阅读 热度 88
星链马斯克乌克兰
加载中...
已经到底啦