Tuesday 27 October

The 60-second version

Five things from this day's batch, best first — everything below the fold expands on them.

1. 1

   New method computes black hole spin tilts at infinity

   A hybrid evolution technique calculates the spin tilts of eccentric, precessing binary black holes at formation, showing eccentricity significantly affects these results. arXiv 2505.07238

2. 2

   Magnetar engine fits fifty-six supernovae lightcurves

   Modeling supernovae with a magnetar engine using fifty-six nickel decays showed consistency across samples, suggesting a unified picture for these explosions. arXiv 2604.21759

3. 3

   Quantum key distribution achieves high bit rate

   A protocol using faithful quantum steering achieved a key rate of 410 bits per second over twenty-seven kilometers of fiber while maintaining strong loss tolerance. arXiv 2604.25225

4. 4

   LLM literacy test validates high school student understanding

   The study developed an objective generative AI literacy test and found it effective for assessing the knowledge of high school students. arXiv 2608.25815

5. 5

   Convex analysis bounds relaxation dynamics in chemical networks

   Using singular values and convexity parameters, the method finds trajectory-dependent bounds for the Kullback--Leibler divergence to equilibrium in reaction networks. arXiv 2602.18932

Across the stations

The day, in brief

Did you know that in the study of black holes, researchers are trying to figure out how to accurately describe systems where the black hole is both spinning and tilted? This isn't just theoretical; this spin tilt tells us about how those massive systems formed in the first place.

The challenge is that we often only measure this tilt when the system is in a very specific frequency band of gravitational-wave detectors. So, scientists have developed a hybrid evolution technique to compute these tilts at infinity. They use orbit-averaged evolution for high frequencies and then switch to precession-averaged evolution as the system evolves toward infinity. While this approach works reasonably well for eccentric systems, both the eccentricity and this hybrid method significantly impact those final tilts at infinity, with some results showing they are not good approximations of how these systems actually formed.

We also have some fascinating developments in astrophysics concerning supernovae. A study modeling supernovae powered by magnetars used fifty-six nickel decay to fit the lightcurves of eighty broad-lined type Ic supernovae. The key finding here is that all the data are consistent with this magnetar engine, suggesting a unified picture for these stripped-envelope events. This work establishes strong correlations between magnetic field strength and the period, pointing toward a common origin for these energetic explosions.

On another front, Euclid’s telescope is pushing the limits of what we can see in space. Researchers used low-resolution near-infrared spectra from the Euclid Deep Field North survey to identify thirty-three new ultracool dwarfs. These faint objects span spectral types M7 to T1 and cover distances between fifty and six hundred parsecs. This work validates Euclid's capability to characterize these faint stellar populations, setting a foundation for larger studies of ultracool dwarfs.

What’s perhaps even more counterintuitive is what happens when we look at quantum materials. In twisted bilayer cuprates, researchers have found that when the pairing involves both s-wave and d-wave components with specific relative phases—not just zero or pi—a state can spontaneously break time-reversal and four rotational symmetries. This leads to a topologically nontrivial phase. It suggests that these twisted cuprates might be a more robust platform for chiral topological superconductivity than we previously thought, opening up exciting new avenues in material design.

The week in the estate's ingest

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

2158 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. Competitive exploration in structured environments [arXiv]
2. SlowLab: Benchmarking Autonomous AI Agents on a Year of Experiments They Cannot Undo [arXiv]
3. Regional prediction of hematoma expansion from non-contrast computed tomography in spontaneous intracerebral hemorrhage patients [arXiv]
4. ODIN: Comparison of Source Extractor and Tractor Source Detection for Lyman Alpha Emitting Galaxy Selection at z=3.1 [arXiv]
5. RobotAgentBench: Towards Open-world Robot Autonomy [arXiv]
6. Dream to Assist: Intent-Conditioned World Models for Shared Vehicle Control [arXiv]
7. Heard Enough to Act: Early Tool Calling in Spoken Dialogue Systems [arXiv]
8. Thinking-in-Bets at Test Time Reduces Sycophantic Revisions in LLM Agents [arXiv]
9. AdaTutor: Adaptive Tutoring through Pedagogical Strategy Selection [arXiv]
10. StreamVibe: Proactive Video Agents for Evolving User Intent [arXiv]
11. Video World Models Need Predictable Latents [arXiv]
12. Neutrinos as Quantum Information Carriers? A Gap Analysis Against the DiVincenzo Criteria [arXiv]
13. Radio view of the Peanut cluster: merger-induced stripping and a radio halo shortly after pericentric passage [arXiv]
14. Hierarchical Task Network planning in the Goal Lifecycle [arXiv]
15. Compound Environmental Operating Stress and Cyber-Induced Pump Disruption in Water Distribution Networks: Matched Hydraulic Impacts, Service Continuity, and Sustained Recovery [arXiv]
16. Agent Memory as Recommendation: Recommender-System Design for Long-Term Memory of LLM Agents [arXiv]
17. Benchmarking Obstacle-Conditioned Safety in Household Robot Navigation [arXiv]
18. From Representations to Computational Trajectories: Intrinsic Fingerprinting for Large Language Model Provenance Verification [arXiv]
19. Curriculum-Guided Acquisition of Hierarchical Task Knowledge from LLM-Generated Experience [arXiv]
20. Exploring perceptual straightness in learned visual representations [arXiv]
21. AT-AT: Automatic Tool Optimization from Agent Traces [arXiv]
22. Pro-Active Inference [arXiv]
23. Beyond Deduplication: Sharp Retrieval Limits for Attention with Correlated Distractors [arXiv]
24. Topic-Level Fairness in Retrieval-Augmented Generation [arXiv]
25. Knowing When the Evidence Is Enough: Requirement-Directed Search with Horn-Closure Certificates [arXiv]
26. Good Rankings, Poor Choices? Rethinking Language-Only Vision-Language Model Selection [arXiv]
27. Machine Learning Applications in the Solar Interior and Atmosphere [arXiv]
28. From Speech Tokens to Stolen Voices: Attacker-Chosen-Text Speaker Impersonation via Representation Inversion [arXiv]
29. RSKD: Grounding Semantic Consistency in Driving VLMs via Distillation [arXiv]
30. IsoAct: Geometry-Aware Post-hoc Debiasing via Isometric Actions [arXiv]
31. Gangqinpu: A Multi-Genre, Multi-Task Piano Dataset with Aligned Audio, MIDI, and Musical Scores [arXiv]
32. Rank-one Factored Gradient Descent without Restricted Isometry: Application to Quantum Pure State Tomography on Multi-Qubits [arXiv]
33. Target Detection with Asymmetric Gaussian Classical Correlation [arXiv]
34. Binary Progenitors of Electron-capture Supernovae [arXiv]
35. Movable-Element Multi-Functional RIS-Assisted Hybrid RSMA-NOMA: A Constrained Hybrid Reinforcement Learning Approach [arXiv]
36. Intrinsic Information Rates of Summaries in Simulation-Based Inference [arXiv]
37. Stable Continuous-Time Consistency Distillation: An Empirical Study with a Multistep Extension [arXiv]
38. Safe by Benchmark, Unsafe in Practice: Lessons from 1331 Child-Related AI Incidents [arXiv]
39. Resource-Performance Trade-offs in Intra-Aisle Robotic Picking: A Discrete-Event Simulation Comparison of Shuttle Configurations With and Without an Onboard Tote Lift [arXiv]
40. Sparse Reconstruction by Direct Measurement-to-Field Generation in Function Space [arXiv]
41. Intervene, Calibrate, Discover: Learning Causal Structure from Pixels [arXiv]
42. SearchHorizonBench: A Benchmark for Search Value in AI Agents [arXiv]
43. Gaia DR3 Compact Binary Candidates Constrained by Radial Velocity Measurements [arXiv]
44. VLA-HWBench: Where the Time Goes When VLA Models Run on Edge Hardware [arXiv]
45. Prefetch-Steered Routing: Deterministic Cache-Aware Expert Selection for Offloaded Mixture-of-Experts Decoding [arXiv]
46. CEPO: A General Framework for Improving Prompt Optimization in Multi-Agent Systems [arXiv]
47. Harness-Model Collaboration for Long-Horizon Continual Learning [arXiv]
48. Designing Customer Feedback for Subscription Boxes [arXiv]
49. SKALOS: Learning to Distinguish Similar Skills with Capability Graphs [arXiv]
50. FGD2HF: Frequency-guided Dual-domain Hypergraph Fusion for Semantic Segmentation [arXiv]
51. iPrefBench: A User-Centric Benchmark for Personalized Mobile Use Agents [arXiv]
52. Optimised growth of epitaxial D019 Mn3Ge(0001) thin films via single target sputtering [arXiv]
53. LADDER: Diagnosing Domain-Level Self-Assessment Under Noisy Feedback [arXiv]
54. Discovering Topics from Linear Autoencoders [arXiv]
55. Networked Humanoid Intelligence: Communication, Coordination, Collaboration, and Social Interaction [arXiv]
56. From Sentinel-2 to 1-m Imagery: Real-World times 10 Cross-Sensor Super-Resolution via Visual Autoregressive Modeling [arXiv]
57. SceneAgent: An Agentic Pipeline from 3D and 4D Captures to Simulation Environments [arXiv]
58. DecisionTraceBench: Tracking Decision-Critical Information in Multi-Party Spoken Conversations [arXiv]
59. Query-Driven Learning for Lifted Reasoning [arXiv]
60. Babysitting AI Agents: A User-Centered Taxonomy and Study of Agentic Delegation Failures [arXiv]

Listen

Everything the estate published on 2026-10-27, in the player. Playing from this page is how the edition learns whether anyone did.

• Daily Summary for 2026-10-27
• The day in conversation for 2026-10-27
• Daily Summary for 2026-10-27
• DeepForestSound: a multi-species automatic detector for passive acoustic monitoring in African tropical forests, a case study in Kibale National Park
• How Embeddings Shape Graph Neural Networks: Classical vs Quantum-Oriented Node Representations
• Graph Neural Networks in the Wilson Loop Representation of Abelian Lattice Gauge Theories
• Generalized Wasserstein Flow Matching: Nested Couplings for Point Cloud Generation
• Temporal Fair Division in Multi-Agent Systems: From Precise Alternation Metrics to Scalable Coordination Proxies
• Tabular Foundation Models for Credit Risk Prediction: A Dedicated Review and Empirical Comparison
• Regret-Based (epsilon, delta) -optimal Stopping Criteria for Bayesian Optimization
• Approaching I/O-optimality for Approximate Attention
• Causal Representation Learning for Generalisable Recommendation

Previous editions

• Sunday 25 October
• Saturday 24 October
• Friday 23 October
• Thursday 22 October
• Wednesday 21 October
• Tuesday 20 October
• Monday 19 October