Top Picks
Authors: Chahe, Amirhosein, Naes, Tyler, D'sa, Jovin, Tariq, Faizan M., Bae, Sangjae, Zhou, Lifeng, Isele, David · 来源：arXiv-list:cs.AI, arXiv-list:cs.CV, arXiv-list:cs.RO · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：Autonomous vehicles must safely navigate complex environments where planning-critical agents may be hidden from view. Current approaches often treat all occlusions with uniform conservatism, yielding needlessly defensive driving, or they infer hidden spaces without estimating the impact on the planner. This work bridges the critical gap between perception and planning by enabling Vision-Language Models (VLMs) to iden…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 5 | 5 | 23 |
Authors: Abdaljalil, Samir, Serpedin, Erchin, Kurban, Hasan · 来源：arXiv-list:cs.AI, arXiv-list:cs.CL · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：We introduce ISOSCI, a benchmark of isomorphic cross-domain science problem pairs that separates reasoning ability from domain knowledge retrieval in LLM evaluation. Each pair shares identical logical structure but requires different domain-specific knowledge, enabling controlled attribution of reasoning-mode gains. Across five model pairs spanning four model families, we find that 91.3% of reasoning-mode gains are k…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 5 | 5 | 23 |
Authors: Liu, Zewen · 来源：arXiv-list:cs.AI, arXiv-list:cs.CL, arXiv-list:cs.LG · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：The bias-reliability tradeoff conjectures that LLM evaluation systems are constrained in (gamma, H, CV) space, where evaluator coupling (gamma), strategy diversity (H), and small-sample measurement reliability (CV(N)) cannot be simultaneously optimized at fixed sample size N. Prior evidence rests on n=5 conditions with complete metrics from a single study. We expand the empirical base to 11 conditions, measuring gamm…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 5 | 5 | 23 |
Authors: Sun, Zhaoyan, Zhong, Shan, Wen, Daizhou, Han, Jiaxing, Li, Guoliang, Yan, Ying, Zhang, Peng, Su, Yu · 来源：arXiv-list:cs.AI, arXiv-list:cs.CL, arXiv-list:cs.LG · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：Data science aims to derive actionable insights from heterogeneous raw data, unlocking the value of the massive amounts of data generated in modern society. Automating this process is essential to reducing labor-intensive efforts for data scientists and enabling scalable data-driven applications. Recently, large language model (LLM)-based data agents have emerged as a promising solution to automate data science workf…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 4 | 5 | 22 |
Authors: Cheng, Xiangchen, Jiang, Yunwei, Sun, Jianwen, Li, Zizhen, Li, Chuanhao, Cao, Xiangcheng, Liu, Yihao, Zhang, Fanrui · 来源：arXiv-list:cs.AI, arXiv-list:cs.CL · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：Memory for a long-horizon LLM agent is a contract about what each future decision is allowed to see. The simplest contract appends past observations, tool calls, and reflections to every prompt, which makes prior context easy to access but also turns it into a jumbled mixture in which the effect of any single memory component is hard to isolate. We introduce and instrument an alternative bounded contract: every decis…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 4 | 5 | 22 |
Authors: Zhang, Yue, Chen, Sihan, Huang, Ziwen, Cui, Hanyun, Ji, Kangye, Wang, Zhi · 来源：arXiv-list:cs.AI · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：LLM-based agents have shown strong potential for solving complex multi-step tasks, yet existing performance improvements often rely on either scaling to larger backbone models or task-specific fine-tuning. The former incurs substantial computational costs, while the latter typically generalizes poorly across different tasks. Although prompt-based control is training-free and broadly applicable, existing methods still…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 4 | 5 | 22 |
Authors: Zhao, Xuan, Chiu, Andy, Wang, Gengyu · 来源：arXiv-list:cs.AI, arXiv-list:cs.IR · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：Information localization within massive repositories is a cornerstone of agentic LLM systems. While synthetic data-driven optimization has proven successful in training LLMs, little attention has been paid to optimizing the agent's working environment (the repository itself) in a data-driven manner. To bridge this gap, we present Libra, a self-evolving framework that introduces mutable "catalogs" (hierarchical Markdo…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 4 | 5 | 22 |
Authors: Abdoli, Sajjad, Al-Sumaidaee, Ghassan, ElShiekh, Ahmad, Taylor, Clayton W., Rashad, Ahmed · 来源：arXiv-list:cs.CL · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：The cost of human expert evaluation is a principal bottleneck to deploying language models in specialized, high-stakes domains. This is particularly acute for Arabic sociolinguistic knowledge: credible grading requires not only linguistic fluency but deep cultural familiarity that cannot be approximated by surface-level metrics. We address this with a cross-evaluation framework instantiated on two underrepresented Ar…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 4 | 5 | 22 |
Authors: Hudson, Blair · 来源：arXiv-list:cs.AI · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：Public LLM leaderboards optimise for global average performance and do not capture the specific cognitive demands of financial-services work: a model that leads on MMLU-Pro may underperform on document-grounded compliance reasoning, and a coding leader may handle multi-turn customer interactions poorly. We present a meta-benchmarking framework that organises 452 publicly reported benchmarks into 41 O*NET Generalized …
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 3 | 5 | 21 |
Authors: Li, Fangfei, Zhao, Chenyang, Wang, Long, Tian, Feng, Zheng, Zhiyue, Guo, Lv · 来源：arXiv-list:cs.AI · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：Domain agents often face noisy business data, uncertain post-training gains, offline/application mismatch, and adapter-release risk. This paper presents CLAP (Closed-Loop Agent Post-training), a closed-loop method that converts business data into structured SFT samples, decision-preference samples, holdout sets, risk diagnostics, and release-gate records. CLAP combines data validation, target/evidence normalization, …
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 3 | 4 | 4 | 5 | 21 |
Authors: Pedinotti, Paolo, Santus, Enrico · 来源：arXiv-list:cs.IR · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：The rapid growth of scientific publications makes it increasingly difficult to track and synthesize research progress. While Large Language Models (LLMs) can support automated survey generation, existing methods retrieve unstructured data and require models to infer conceptual, methodological, and taxonomic relations from raw text at generation time. We introduce STRUCTSURVEY, a hierarchical multi-agent framework tha…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 3 | 5 | 21 |
Authors: Tagliapietra, Nicholas, Marchioni, Gian Lorenzo, Willig, Moritz, Luettin, Juergen, Halilaj, Lavdim, Kersting, Kristian · 来源：arXiv-list:cs.AI, arXiv-list:cs.MA · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：Learning causal models from high-dimensional data is a significant challenge, particularly in real-world settings where violations of core assumptions lead to causal identifiability issues. Although massive amounts of prior knowledge are available, and contain valuable causal information, effectively integrating this knowledge into the causal discovery process remains an open problem. We introduce CausalSTeward (CAST…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 3 | 5 | 21 |
Authors: Le, Cuong Chi, Yadavally, Aashish, Le-Anh, Minh, Nguyen, Tien N. · 来源：arXiv-list:cs.SE · PDF
一句话结论：提供新的评测/基准线索，可用于校准 Agent 或 LLM 系统能力边界。
方法要点：Large language models (LLMs) are typically evaluated on code generation and program repair using binary functional correctness: a generated program or patch either passes or fails a test suite. This protocol is simple but coarse, as it ignores partial progress, feedback use, regressions, and the refinement trajectory through which models often improve code. We introduce PAIR-Bench, a progressive and adaptive benchmar…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 3 | 5 | 21 |
Authors: Liu, Tianci, Dong, Zihan, Zhang, Linjun, Wang, Haoyu, Gao, jing, Kiciman, Emre, Chandra, Ranveer, Chen, Wei-Ting · 来源：arXiv-list:cs.AI · PDF
一句话结论：直接关联 Agent 架构、长程任务或工具使用，是后续产品/研究原型的高价值输入。
方法要点：Slide design requires personalizing both deck themes and page layouts. Yet, current AI agent-based methods struggle with fine-grained, page-level design. Solely relying on prespecified templates or user verbose instructions, they fail to capture latent design intents, leaving Page-level Slide Personalization (PSP) unresolved. To close this gap, this work formulates PSP as an inverse planning problem. We propose to le…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 4 | 2 | 5 | 20 |
Authors: Li, Jisen, Li, Bingxuan, Jiang, Nanyi, Ning, Xuying, Wang, Xiyao, Shen, Yifan, Wang, Heng, Jian, Yuqing · 来源：arXiv-list:cs.CL, arXiv-list:cs.IR · PDF
一句话结论：直接关联 Agent 架构、长程任务或工具使用，是后续产品/研究原型的高价值输入。
方法要点：Scientific literature search often requires more than retrieving papers from a single query: users' intents are underspecified, preference-dependent, and evolve through interaction. Existing search agents typically rely on fixed pipelines or implicit language-only reasoning, making their search strategies difficult to control, inspect, and refine. We introduce PaperPilot, a multi-turn literature search agent that fra…
Lucian 下一步：抽取任务定义、评测脚本和失败案例，判断是否可转成 Auto Research 实验卡。
| Relevance | Novelty | Substance | Evidence | Actionability | Total |
|---|
| 5 | 4 | 3 | 3 | 5 | 20 |