Wednesday 16 September

The 60-second version

Today's batch, and the 60-second version is still being written. Everything below is today's own work: the papers as they are landing, the stations that have written their piece, and the day's episodes. This page fills in as the day does.

The week in the estate's ingest

Seven days of papers the estate ingested. Weekends are striped — the weekend is visible in the data.

676 papers, three ways

Where today's batch landed

The index

The day's strongest 60, by score. Every title links its paper here and on arXiv.

1. Acoustic and perceptual differences between standard and accented speech and their voice clones [arXiv] [episode]
2. RoofSeg: An edge-aware transformer-based network for end-to-end roof plane segmentation [arXiv] [episode]
3. Shared Selective Persistent Memory for Agentic LLM Systems [arXiv] [episode]
4. Comparative Characterization of KV Cache Management Strategies for LLM Inference [arXiv] [episode]
5. The Verifier is the Curriculum: Precision Sets the Return on Search in Code Self-Distillation [arXiv] [episode]
6. Rhythm of the Deep: Two-tier acoustic organization of sperm-whale codas from click waveforms to second-order sequence dependence [arXiv] [episode]
7. Scalable Algorithms for Approximate DNF Model Counting [arXiv] [episode]
8. Ave: Guiding Agentic GPU Optimization Using Data-Flow Invariants [arXiv] [episode]
9. FactorEngine: A Program-level Knowledge-Infused Factor Mining Framework for Quantitative Investment [arXiv] [episode]
10. Can LLMs Model Incorrect Student Reasoning? A Case Study on Distractor Generation [arXiv] [episode]
11. MASCOT: Multi-Agent Socio-Collaborative Companion Systems [arXiv] [episode]
12. Composable multi-satellite precipitation estimation for evolving observing systems [arXiv] [episode]
13. Liberating LLM Capabilities in Full-Duplex Speech Models [arXiv] [episode]
14. Measuring Curriculum Alignment across Topical Coverage, Competency, and Cognitive Depth: A Longitudinal Framework Applied to CS2013 and CS2023 [arXiv] [episode]
15. Multi-Agent Collaboration for Automated Design Exploration on High Performance Computing Systems [arXiv] [episode]
16. T2T-VICL: Cross-Task Visual In-Context Learning via Implicit Text-Driven VLMs [arXiv] [episode]
17. MiCRo: Mixture Modeling and Context-aware Routing for Personalized Preference Learning [arXiv] [episode]
18. Lost at the End: Primacy Bias in Multimodal Retrieval-Augmented Question Answering [arXiv] [episode]
19. LLM-Guided Dynamic Action Spaces for Synthesizable Molecular Optimization [arXiv] [episode]
20. Resolution scaling governs DINOv3 transfer performance in chest radiograph classification [arXiv] [episode]
21. DoubleAgents: Human-Agent Alignment in a Socially Embedded Workflow [arXiv] [episode]
22. CASCADE: An Agentic Regulatory Network Framework for Patient-Data-Validated Downstream Perturbation Prediction [arXiv] [episode]
23. Hierarchical Modeling of ICD Codes in EHR Foundation Models [arXiv] [episode]
24. Robustness as an Emergent Property of Task Performance [arXiv] [episode]
25. Autonomous Assessment of Generalizability of AI Agent Capabilities [arXiv] [episode]
26. Dual Randomized Smoothing: Beyond Global Noise Variance [arXiv] [episode]
27. Say the Mission, Execute the Swarm: Agent-Enhanced LLM Reasoning in the Web-of-Drones [arXiv] [episode]
28. Graph Neural Assisted Actor-Critic for Latency-Efficient Edge Vision System [arXiv] [episode]
29. Superficial Beliefs in LLM Decision-Making [arXiv] [episode]
30. Generating Individual Travel Diaries Using Large Language Models Informed by Census and Land-Use Data [arXiv] [episode]
31. The Biomimetic Architecture of Software 4.0 [arXiv] [episode]
32. Adaptive Bayesian Partner Selection for Federated Clinical Centers [arXiv]
33. Charging of rotating black holes: kinetic simulations of black hole magnetospheres [arXiv]
34. Decentralized Gossip Learning and Federated Averaging for Histopathology Image Classification [arXiv]
35. Early-Bird Decoding: Accelerating Diffusion LLMs with Learnable Block Sizes and Parallel Sampling [arXiv]
36. Fine-Tuning Fixes Mode Collapse and Over-Dispersion in LLMs [arXiv]
37. The Variation of Circumstellar Parameters through Early Stellar Evolution [arXiv]
38. Language Orthogonalization for Zero-Shot Cross-Lingual Audio Deepfake Detection [arXiv]
39. OPD-Aha: From Linguistic Momentum to Visual Reflection in Multimodal On-Policy Distillation [arXiv]
40. Protocol-Preserving Context Trimming for Agentic Workflows: Benefits, Failure Regimes, and Budget Guardrails [arXiv]
41. Not All Relations Are Equal: Relation-Balanced and Calibrated Graph Learning for Provenance-Based Intrusion Detection [arXiv]
42. A multimodal large language model for evidence-based autism spectrum disorder screening [arXiv]
43. Online Gradient Computation for Warping Gaussian Process Transformations [arXiv]
44. The Study of Detailed Morphology of Milky Way Dwarf Galaxies: Draco and Bo"otes I [arXiv]
45. Certified Inference and Training for Deep Equilibrium Networks: A Continuation Framework with Polynomial Complexity Guarantees [arXiv]
46. Skill-based Agentic Evaluation for Real-time Data Science Tasks [arXiv]
47. Decoder Design Matters for ECG Delineation [arXiv]
48. From Manual Construction to AI-Driven Scenario Emergence: Rethinking Catastrophe Risk Modeling [arXiv]
49. AI Policies: Help or Hindrance? A Software Developer's Perspective [arXiv]
50. Geospatial Metadata Improves Discoverability by Connecting Datasets Across Scientific Disciplines [arXiv]
51. High-Performance Tensor Formulation of the Viterbi Algorithm for Hidden Semi-Markov Models [arXiv]
52. Beyond the Name: Demographic Leakage in De-Identified R'esum'es and Evaluation Artifacts in LLM Bias Audits [arXiv]
53. Causal Path Analysis from Perturbational and Population-Scale Single-Cell Data with Multiscale Confounding and Measurement Error [arXiv]
54. Competence-Preserving Resume Perturbations Expose Presentation Sensitivity in LLM Screening [arXiv]
55. AquiLLM: Evaluating Faithfulness in Open-Weight RAG-LLM Systems for Scientific Research [arXiv]
56. QALPA: Property-guided diffusion modeling for efficient exploration of chemical spaces of flexible molecules [arXiv]
57. FlowATC: Aircraft Trajectory Prediction via Flow Matching [arXiv]
58. Style-Debiased DPO: Updating LLM Knowledge with Factuality-Aware Synthetic Preference Data [arXiv]
59. Bridging Between Statistical Mechanics and Black Hole Evolution: Theoretical Formalism [arXiv]
60. What Does Layer-Importance Reveal About Transformers and State-Space Models? [arXiv]

Listen

Everything the estate published on 2026-09-16, in the player. Playing from this page is how the edition learns whether anyone did.

• FactorEngine: A Program-level Knowledge-Infused Factor Mining Framework for Quantitative Investment
• LLM-Guided Dynamic Action Spaces for Synthesizable Molecular Optimization
• Comparative Characterization of KV Cache Management Strategies for LLM Inference
• Ave: Guiding Agentic GPU Optimization Using Data-Flow Invariants
• Say the Mission, Execute the Swarm: Agent-Enhanced LLM Reasoning in the Web-of-Drones
• Moirae: A Multimodal Agent Collaborative Framework for Dynamic Android Malware Detection
• Text-Driven Artistic Staging: 3D Posing, Lighting, and Camera References from Paintings
• Liberating LLM Capabilities in Full-Duplex Speech Models
• The Biomimetic Architecture of Software 4.0
• Composable multi-satellite precipitation estimation for evolving observing systems
• Hierarchical Modeling of ICD Codes in EHR Foundation Models
• Superficial Beliefs in LLM Decision-Making