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.
How Children Design and Reason about Trustworthy AI Chatbots
Deniz Ozturk, Jiayu Li, Daksh Pratap Singh, Yasitha Rajapaksha et al.
Valuable mixed-methods study of children as chatbot designers; 7 trust dimensions credible, but age/domain effects are confounded and exploratory.
Geometric and Semantic Coupling for Interaction Understanding in 3D Scenes
Hanyang Kong, Xingyi Yang
Claims handle-guided hinge selection lifts motion AP 13.74→40.98 on Articulate3D; rigorously ablated and honest, but single-dataset and validation-tuned.
The AI Neuroscientist: An Interactive Agentic Interface for Neuroimaging Analysis
Aakash Patel, Panos Ketonis, Shreya Saxena, Smita Krishnaswamy et al.
fNIRS tool-agent beats sandbox LLMs on a self-built 6-task benchmark (0.79 vs 0.45); convincing demo, but benchmark circularity and no code release temper it.
Lean Pool: An AI-Maintained Archive of Formalized Mathematics
Vasily Ilin
AI-maintained archive of 211 Lean formalizations (3.23M lines) with operational records, but review accuracy and reuse claims lack independent validation.
MedGate-Fusion: Integrating First-Encounter Semantic Narratives and Physiological Biomarkers for Prospective Stroke Risk Stratification
Hemn Khdr, Mohammad Noaeen, Karim Keshavjee, Aziz Guergachi et al.
5-year stroke prediction from first-visit text+10 biomarkers (AUROC 0.848, fusion); text dominates; residual leakage and missing fusion ablation temper claims.
When LLM Agents Fail to Read the Room: ReAdapt for Relational Social Reasoning
Jianzhe Lin, Xiaolin Li, Yunda Liu, Fei Wang et al.
Claims explicit social-state updates lift LLM-agent relational accuracy (37→51%, 69→77%); credible proof-of-concept but single model, no ablation/stats.
FineWeb-CLaR: Culture, Language, and Region Annotations for Benchmark-Aligned Corpus Auditing
Yusser Al Ghussin, Eva Gavaller, Cristina España-Bonet, Josef van Genabith et al.
Culture-region annotation of 30.9B FineWeb docs; candid about weak CT-label validity (22% human agreement); valuable auditing resource, exploratory analyses.
Attention as a Routing Graph: Live Circuit Extraction from a Single Forward Pass
Ash Manvi, Samreena Tajreen
One-forward-pass attention routing sketch shows real causal signal on induction/IOI, but uniform random controls confound attention mass with routing effect.
Learned Enterprise Data Comprehension: Compression and Routing for Data Agents
Ethan Torres, Eric Mills
Lifts DAB Pass@1 to 94.67% (Core 84.13%) via prototype routing; promising, but confounded ablation, undisclosed evidence pipeline, closed code.
Correcting Within-Group Self-Selection Bias in Prioritized Replay
Oscar Miró López-Feliu, Herke van Hoof
Proposes sibling-aware replay to fix PER's within-group outcome bias; theory is clean, evaluation honest, but gains over plain uniform replay stay unproven.
Topological Signal Processing With Unoriented Operators
Andrea Cavallo, Varun Sarathchandran, Geert Leus, Elvin Isufi
Proposes an unoriented TSP framework with a proven interaction-order decomposition; theory is sound and regularizers win on four small real datasets.
Spatiotemporal Kronecker Covariance Neural Networks
Andrea Cavallo, Athanasios Georgoutsos, Elvin Isufi
KVNN decouples spatiotemporal covariance via Kronecker sums; theory is sound and forecasting parameter-efficient, but baselines are dated and margins modest.
Making Agents More Consistent: Skills Should Form Habits for Repeat Tasks
Travis Weber, Rohit Taneja
Proposes coverage-gated skill habituation for consistent, conditionally cheaper repeated agent work; candid failure analysis, but self-authored evals, small n.
Potential for Enhanced Learning in Machine Learning Classes by Using Wiki LLM Indexing
Brian Wright
Small single-course study: an LLM-compiled wiki beats vector RAG on accuracy and groundedness; promising but confounded, with no learning outcomes measured.
MT-ProtBERT: Multi-task Learning ProtBERT for Intrinsically Disordered Proteins Classification with Scarce Data
Jian Sun, Kingshuk Ghosh, Lilianna Houston, Mohammad H. Mahoor
Multi-task ProtBERT for scarce-data IDP classification beats PARROT; results directionally credible but tables conflict, baselines implausible, no code.
Clarification Is Not Correction: LLMs Fail to Let Go
Jianzhe Lin, Xiaolin Li, Fei Wang, Robert Douglas et al.
Claims LLMs commit early and treat later clarification as context, not correction; compelling diagnosis, but judge circularity and no stats tests weaken it.
GINIO: A Geometric SO(3)-Equivariant Interface for Neural Inertial Odometry
Chankyo Kim, Minghan Zhu, Tzu-Yuan Lin, Avantika Rattan et al.
Equivariant NIO interface, provable LFA equivalence; rotated-mount gains convincing (TLIO 2.0 vs 76.4 m), nominal wins modest, code only promised.
From Decorative to Load-Bearing: Task Difficulty Shapes the Causal Role of Chain-of-Thought
Renee Jia, Di Mu
Claims CoT turns load-bearing as task difficulty rises; unusually rigorous self-validation, but headline BBH C-rates rest on partially miscalibrated labels.
Mean Velocity Matching: Rethinking Generative Dynamics in Diffusion Models
Yunhong Zhang, Changjie Cao, Zhihua Zhang, Bingli Liu et al.
Proposes single-field drift for a Gaussian diffusion path with correct theory and strong FIDs, but novelty is algebraic and key ablations are missing.
TelecomGPT-R1: Unified Post-Training for Reasoning Across Heterogeneous Telecom Tasks
Bohao Wang, Chenwei Wu, Hang Zou, Yu Tian et al.
Unified telecom reasoner claims 89.6% mean on GSMA Open Telco, beating GPT-5 and Claude; solid ablations, but contamination and eval-setting caveats persist.
Sex Estimation from Footwear Outsole Impressions Using CNN Transfer Learning and Interpretable Image Statistics
Jinyi Niu, Ziyi Song, Weining Shen
声称CNN迁移学习可通过鞋底图像估算性别，测试准确率高达100%；设计合理，但在单一数据划分上获得完美分数值得谨慎对待。
SSP-Bench: A Hybrid Data Generation Framework for Safety, Security, and Privacy Evaluation
Fatih Deniz, Yazan Boshmaf, Issa Khalil
Dynamic SSP benchmark exposing static-leaderboard failures via construct mixing; strong τ evidence but authors' own baseline and partly n.s. moderation leg.
Concept Drift from a Causal Perspective
Eduardo V. L. Barboza, Jean Paul Barddal, Robert Sabourin, Rafael M. O. Cruz
Introduces a 5-type SCM causal taxonomy of concept drift plus the CaDrift generator; well-grounded theory, but stats rigor and augmentation baselines are thin.
Learning from Humans for Proactive Assistance in Human-Robot Collaborative Transport
Elvin Yang, Christoforos Mavrogiannis
PROACT pairs learned human-behavior prediction with compliant control; 108 real-robot trials show large effort reductions, but statistical evidence is thin.