Evolve Quick Reviews
Public Paper Reviews
Structured preliminary reviews of recent open-access papers, generated and continuously updated by the Evolve analysis engine. Each review condenses what the paper claims, how solid the evidence is, and the questions a rigorous peer reviewer would ask — English analysis with abstracts in your language. These are independent AI analyses, not peer review; the canonical record always lives at the publisher.
Think Locally, Refine Globally for Memory-Efficient 3D Reconstruction
Jingke Zhou, Chenhang Ma, Zhizhou Zhong, Mingkai Liu et al.
LoG-VGGT: windowed cross-attention + register-based global pose refinement for bounded-memory long-sequence 3D reconstruction; gains real, no variance or code.
A Cyber Range Evaluation of Autonomous Network Incident Response Agents
Jakob Nyberg, Teodor Sommestad, Andrei Buhaiu, Joakim Loxdal et al.
Claims RL agents beat a heuristic in a 37-VM cyber range; credible sim-to-real study, but stats are thin and NoOp wins under Guided attacker.
PunGraph: Retrieval-Enhanced Phonetic-Semantic Graph Reasoning for Pun Understanding
Yuchen Su, Zijian Huang, Yaotian Shi, Shaoxin Zhong et al.
PunGraph uses phonetic-semantic graph retrieval to boost LLM pun reasoning; gains are large for 7B models but shrink with scale, and thresholds risk test leakage.
Large Language Models in the Loop: A Stability- and Network-Aware Survey in Networked Control, Cyber-Physical, and Multi-Agent Systems
Haiping Du, Linping Chan
A survey casting LLMs as slow supervisors over certified control loops; conceptually strong and honest, but its ISS stability premise is an unproven template.
Supervising the Chain Ladder
Stephan Marais, James Grove
Refrees actuarial pattern adjustments as penalties on the chain-ladder loss with holdout training; sound math, but evidence is one hand-picked triangle.
QueryFormer: Winning Solution for KDD Cup 2026 Tencent UniRec Challenge
Yuanzhe Zhou, Zhaoyang Zeng
KDD Cup 2026 winner with cross-attention query generation; strong industrial evidence, but 1e-4-scale AUC effects lack variance or significance testing.
The MAL Simulator: Cyber Operations Simulation based on Attack & Defense Graphs
Jakob Nyberg, Sandor Berglund, Andrei Buhaiu, Joakim Loxdal et al.
The MAL Simulator adds language-driven adaptability and dual attack/defense interfaces; case-study RL results are credible but validated only in simulation.
Query-Aware Source-Risk Triage for Retrieval-Augmented Generation
Kainan Zhou, Gangzhen Qian, Chuhong Xu, Lu Yi
Proposes query-aware source-risk triage for RAG with auditable scores; honest about limits but evidence is a single-coded pilot plus synthetic scenario only.
GameReplica: A Benchmark for Black-Box Visual Game Replication by Vision-Language Agents
Boyu Qiao, Zixin Tang, Xiaoshuai Hao, Wenbo Li
Claims first black-box game-replication benchmark; isolation is well enforced, but LLM-judge scoring, no significance tests, and contamination risks remain.
Recovering Physical Parameters from Fragmented Observations via Exact Distributed Spline Merging
Naveen Mysore
Exact one-shot merge of ridge spline statistics — a classical identity, honestly scoped; thin novelty, physical recovery shown only on synthetic PDEs.
Stable by Construction: Variational Latent Markov Operators for Long-Horizon PDE Prediction
Junyi Liao, Johann Guilleminot, Vahid Tarokh
VAMO adds variational latent Markov dynamics to neural operators; strong long-horizon fluid results, theory-practice aligned, but only self-built baselines.
CLASH: Counterfactual Auditing of Lexical and Prosodic Reliance in Spoken Sarcasm Detection
Qiyang Sun, Xudong Li, Yupei Li, Jiabin Xue et al.
Careful counterfactual audit: Qwen3-Omni's lexical-over-prosody AUROC edge survives duration balancing and TTS resynthesis; rigorous but model-specific.
Do LLMs Have Values? A Quantitative Analysis and Alignment Framework for Values in Large Language Models
Keqing Zhang, Jingyu Chen, Yufan Liu, Yongqiang Zhu et al.
Claims LLMs crystallize into a narrow idealized value core, quantified via a new PEC framework; impressive scale, but LLM-based item labels risk circularity.
A Framework for Generating Valid Context-Specific Benchmarks through Expert Guidance
Kimberly Le Truong, Nari Johnson, Anna Kawakami, Hoda Heidari
Schema-guided synthetic benchmark generation with expert input; validity metrics + case study show realism/coverage gains, but evidence is single-domain.
Challenges of Auditing: Variability in Outputs of Large Language Models for Health
Yuan Pu, Yewon Chang, Furong Jia, Xunjian Yin et al.
Finds systematic ChatGPT-vs-API output differences for health queries; sound paired design, but uncorrected p-values and LLM-judged features temper claims.
AI in Science: Early Insights
Mihai Codreanu, Alex Imas, Juan Mateos-Garcia, Joseph Emmens et al.
Google-led study of Gemini logs, 2,690 specialist models and 637 scientists finds broad AI adoption and complementarity; rich but descriptive, vendor-tilted.
A Weighted Kernel Method for Approximation that Adapts to Learned Multivariable Structure
John E. Darges, Laura Weidensager
Weighted ANOVA kernels with d learned factors; strong-convexity and consistency proven, but no rates and single-run experiments. Partially validated.
Structure Across Voices: Comparing acoustic-event type accumulation and sequence dependence across four vocal repertoires using frozen audio encoders
Mudit Sinha, Sanika Chavan
Claims four vocal repertoires differ by acoustic property and scale, not one hierarchy; careful matched-null design, but whale sample (3 units) limits reach.
DiaWhisper-DPO: Role-Attributed Transcription of Clinical Interviews via Failure-Mined Preference Optimization
Weiming Li, Ana Catarina Fidalgo Barata, Miguel Constante, João Miguel Sanches
Whisper + failure-mined DPO for clinician/patient-tagged transcription; impressive 0.973 role accuracy but on synthetic DAIC audio, vs cascaded-only baselines.
RoleBreak: Benchmarking Long-Horizon Role-Playing Robustness in Spoken Dialogue
Yuqi Wang, Fengyuan Liu, Haochen Luo, Zhiqi Yu et al.
Benchmark of 310 roles shows spoken role-playing breaks by ~turn 10; valuable resource, but single LLM judge and no error bars temper the headline claims.
Divergence Timing and Cumulative Disagreement under KV-Cache Eviction
Xinyue Luo, Fei Yu
Derives exact divergence-timing vs exposure decomposition of KV-eviction disagreement; theory is sharp and stats honest, but only 2 models, H=128, no code.
Characterizing Heterogeneous Rates in Finite Mixture Estimation via Partial Optimal Transport
Dung Le, Huy Nguyen, Trang Pham, Alessandro Rinaldo et al.
VPOT: Voronoi partial-OT loss giving configuration-adaptive minimax n^-1/2 MLE rates for mixtures in any dimension; rigorous theory, no experiments.
AURA: Agentic Diagnosis and Refinement for Production Recommender Systems at Scale
SungGeun Kim, Abhinav Narain, Daniel Nemirovsky
Disney's AURA: LLM agents diagnose recsys failures from production logs and emit code fixes; honest negatives, but evaluation is LLM-judged, no human audit.
EchoPath: Execution-Level Replayable Memory for GUI Agents
Yao Zhao, Aditya Shanmugham, Swastik Roy, Yanxun Xu
GUI-trajectory memories replayed via image re-aiming cut tokens >90% and time ~60% on OSWorld reruns; solid systems work but only same-task reuse is tested.