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.
General Collaborative Intelligence: Architecting Cognition for Resilient Multi-Agent Ecosystems
Lei Zhang, Chun Ye, Le Yang, Zhaozhong Wang et al.
Survey of collaborative perception via 5-D taxonomy, 3 synergy conditions and GCI-Bench; strong synthesis, but case-study scores are illustrative only.
When Agentic Trust Crosses Organizational Boundaries: Structural Externalization and a Reference Model for Trust Evidence
Huafu Li, Jia Xia
Conceptual paper proposes a diagnostic, trust-evidence envelope, and eval protocol for cross-domain agentic trust; carefully scoped yet unvalidated.
A Horizon-slicing Approach to Minimum Obstacle Displacement Planning for Robot Navigation
Antony Thomas, Giulio Ferro, Fulvio Mastrogiovanni, Michela Robba et al.
Proves MOD NP-hard for 3 obstacle classes; exact MIQP + horizon-slicing heuristic. Sound and honest, but gaps up to ε≈4.7 and thin baselines warrant caution.
Beyond Similarity: Coverage-Aware Prompt Selection for Time Series Forecasting with LLMs
Daeun Ji, Minkyoung Kim, Dongkuk Kim, Yohan Lee et al.
Parameter-free coverage regularizer for LLM prompt retrieval in forecasting; unusually honest stats, but the aggregate gain resolves on only 1 of 24 cells.
Dual-Locking Learned AI Models: A PIN-Based Sparse QIM Watermarking and Adaptive Index Permutation Approach
Iva Vasic, Jesús Muñoz-Cádiz, Bata Vasic
Claims reversible weight-permutation locking + QIM bias watermarking secure AI models; promising results, but evaluation and security analysis lack rigor.
Connectivity-Aware Exploration of Robotic Grasp Spaces
Maksim A Kazanskii
Credible simulation-only study: connectivity-aware grasp-set sampling speeds structure reconstruction (2-5x) but adds little over local viability scoring.
LPINNs: First-Layer Gated Localization for Physics-Informed Neural Networks
Lakshay Chawla, Hardik Jain
PINN first-layer gating cuts error 1-2 orders on three toy benchmarks; clean paired-seed evidence, but post-hoc config choice and only a vanilla baseline.
Interpretable Multi-Hypersphere Deep Anomaly Detection for Open-set Supervised Anomaly Detection
Zhiji Yang, Fangyong Wang, Yue Li, Xianli Pan et al.
Per-class hyperspheres embedded in the final layer give interpretable open-set anomaly detection; strong AUC wins, but baselines are dated and single-seed.
Rethinking Pivot Programming Languages in Code Language Models
Andor Diera, Lukas Galke Poech, Matthias Tichy
Claims no universal pivot PL: geometry+retrieval show relation-dependent structure (Python=English-facing); well-controlled, modest scale.
OptiSkill: A Hierarchical and Evolving SkillBank for LLM-Based Optimization Modeling
Ruiqing Zhao, Rui Liu, Yuan Zuo, Huarong Zhang et al.
Solver-verified hierarchical SkillBank tops OR-modeling agentic baselines (73.9% avg, DeepSeek-V4); convincing, but no significance tests and only 2 backbones.
On attention heads and bilinear forms
Andrew O'Desky
Proves attention-head QK symmetric parts are balanced (n,n,N−2n) via a new profile map; theorems sound, but learned spectral patterns stay unexplained.
Reconstructed holograms and explanation-aware evaluation for low-cost computational pollen analysis in veterinary cytology
Swarn Warshaneyan, Joial Danyal, Blaž Cugmas, Mindaugas Tamošiūnas et al.
Hologram reconstruction cuts the brightfield-DIHM pollen-analysis gap substantially; XAI diagnosis is credible but rests on six test images and private data.
Auditing Political Alignment in LLM Assistants: Engagement, Stance, and User Identity
Joan C. Timoneda
Claims LLM political speech follows user-conditional 'speech regimes'; a preregistered 7,500-conversation audit is persuasive but rests on LLM judges.
Silent Failures Beyond the 32-Bit Index Range: A Differential Characterization of Large-Tensor Matrix Multiplication in PyTorch's MPS Backend
Junichiro Niimi
Rigorous black-box study finding silent MPS bmm corruption above 2^32 elements in PyTorch across 11 releases; systematic, reproducible, convincing.
Spatial-Interactor: Learning Spatial Reasoning through Interaction with the Observable Physical World
Kaixiang Yao, Xu Wang, Miao Pan, Hu Xiyue et al.
Interaction-trajectory training (LSI-108K; SFT+OPD) lifts VLM spatial reasoning broadly, but its own ablation shows generic spatial-QA SFT drives most gains.
Watching Quantum Models Think: Hilbert-Space Interpretability in Quantum Transformer Blocks
Diego Iacopetta, Andrea Gasparini
Proposes mutual-information tracking as intrinsic interpretability for a 12-qubit Quantum Transformer Block; mostly convincing proof of concept, but key circuit details are IP-withheld.
PINNForge: Execution-Grounded Evolutionary Design of Physics-Informed Neural Networks for PDE Solving via Large Language Models
Mingyang Yu, Xu Yang, Jun Zhang, Xiaolong Wang et al.
PINNsForge: LLM execution-feedback evolution wins 24/25 PDE tasks with solid stats; credible, but novelty vs Lang-PINN and uncounted LLM cost temper the claim.
WaveFront Decoding: Parallelized Self-Speculative Decoding for Looped Language Models
Hyeongju Ha, Jae-Joon Kim
Credible training-free co-batched draft-verify decoding for looped LMs with solid speedups; evaluation is narrow — two checkpoints, one GPU, greedy-only.
Enforcing Narrative Reliability and Epistemic Pacing in LLM-Driven Detective Games via Structured Knowledge Trees
Parsa Rahmati, Richard Zhao
Proposes SKT+tri-agent pipeline: critical hallucinations 17.80%→6.27%, premature confessions 10/33→0/33; promising but confounded evaluation, no code.
Anatomy of a Closed-Loop Collapse: A Causal Case Study of a Compressed VLA Policy
Fengze Jia
Claims a causal anatomy of one compression-induced closed-loop collapse: four therapies fail, teacher-rollout substitution suffices. Rigorous but single-run, n=1.
Optimizers for Diffusion Models: A Controlled Benchmark
Arman Bolatov, Egor Shulgin, David Li, Abduragim Shtanchaev et al.
First controlled optimizer benchmark for discrete diffusion; well-run and reproducible, but text-task budgets are just 3% of published training length.
From Concept Alignment to Causal Grounding: An Intervention Test of Chain-of-Thought Faithfulness
Qianli Wang, Yilong Wang, Dennis Wei, Jingyi Sun et al.
Reframes CoT faithfulness as SAE concept grounding with a validated causal metric Δp across 5 LLMs; strong validation, but conceptual and scope caveats remain.
Attributable Post-Rationalization in RAG Citations: A Controlled Reproduction and an RLVR Comparison
Mehedi Khan, Md. Shariful Islam Bhuyan
Claims answer-only RLVR leaves RAG citation faithfulness unchanged; controlled reproduction is persuasive but lacks human validation and an SFT control.
Bridging Static and Agentic RAG for Taiwanese Historical Question Answering
Kai-Hsin Chen, Wei-Yu Chen, Xuanjun Chen, Jyh-Shing Roger Jang
Agentic vs static RAG tie on aggregate yet differ on 70.8% of questions; post-hoc selection recovers 60% of oracle headroom — solid but LLM-judge dependent.