AI/ML arXiv cs.AI

Simulating Eutopia: Revisiting Long-term Fairness with Outcomes, Performativity, and Dynamics

The paper introduces Eutopia, a lending-process simulator used to study and improve long-term fairness and equity in AI-driven decision makers.

AI/ML arXiv cs.AI

LAARA: Layer-Aware Adaptive Rank Allocation for Parameter-Efficient Fine-Tuning

LAARA is a search-free framework for parameter-efficient fine-tuning that dynamically allocates ranks to transformer layers using Fisher estimates.

AI/ML arXiv cs.AI

Decodable but Not Detectable: A Leakage Fingerprint for Near-OOD Benchmarks

The authors identify a 'leakage fingerprint' to detect when OOD benchmarks are contaminated with training data, which can lead to misleadingly high performance scores.

AI/ML arXiv cs.AI

Cross-Subject Semantic Decoding with Shared-Space Alignment for Generalized Neural Representation Learning

A new framework for cross-subject semantic decoding aligns neural responses to speech perception into a shared latent space to improve generalization across different individuals.

AI/ML arXiv cs.AI

From Trajectories to Prefixes: Reusing Teacher Trajectories via Replayed Prefixes and Online Continuation

Prefix-GRPO is a reinforcement learning framework that improves small-model agents by decomposing teacher trajectories into replayed prefixes and online continuations.

AI/ML arXiv cs.AI

Leveraging Offline Supervision for Efficient and Generalizable Reinforcement Learning in Large-Scale Vision-Language-Action Models

This work investigates hybrid offline-online training for Vision-Language-Action (VLA) models, showing that offline supervision can double training efficiency while maintaining OOD performance.

Software Engineering Hacker News

Escape IntelliJ: Scala and Kotlin LSPs on Emacs Eglot

Discussion on configuring Scala and Kotlin Language Server Protocol (LSP) support in Emacs using the Eglot package.

Software Engineering Hacker News

Cruller: Bun's Zig Runtime, Continued on Zig 0.16

Updates on Cruller, a Zig-based runtime for Bun, now continuing development on Zig version 0.16.

AI/ML arXiv cs.AI

Global Difference Constraint Propagation for Constraint Programming

A research paper proposing a global propagator for difference constraints in constraint programming to improve solving speed and completeness.

AI/ML arXiv cs.AI

Reading and Steering Representations of Materials-Science Mechanisms in an Open-Weight Language Model

Research on steering representations of materials science mechanisms within the open-weight Gemma-4 model to understand how LLMs represent physical laws.

AI/ML arXiv cs.AI

PRO-LONG: Programmatic Memory Enables Long-Horizon Reasoning

Introduction of PRO-LONG, a context management framework using programmatic memory to enable LLM agents to handle long-horizon reasoning tasks more efficiently.

AI/ML arXiv cs.AI

TRUST-ESD: A Risk-Calibrated and Governance-Aware AI Framework for Enterprise Strategic Decision Support Under Uncertainty

Presentation of TRUST-ESD, a risk-calibrated AI framework designed for enterprise strategic decision support under uncertainty.

AI/ML arXiv cs.AI

CUSUM-Shaped Inference-Time Monitoring and Targeted Re-Decoding for Quantized Small Language Model Reasoning

A method called MGT-B for monitoring and targeted re-decoding in quantized small language models to prevent repetitive or unproductive reasoning trajectories.

AI/ML arXiv cs.AI

PoTRE: Test-Time Reasoning inspired by Cognitive Heterogeneity

Introduction of PoTRE, a heterogeneous ensemble framework that decouples LLM inference into four specialized agents to improve complex reasoning performance.

AI/ML arXiv cs.AI

Train the Model, Not the Reader: Decodability Supervision for Verifiable Activation Explanations

Analysis of activation explanations in LLMs and the introduction of RECAP, a method to make internal content independently checkable via co-trained auxiliary predictors.

AI/ML arXiv cs.AI

SoftReason: A Fully Differentiable Neuro-Soft-Symbolic Deductive Reasoning Architecture over High-Dimensional Perceptual Data

Introduction of SoftReason, a fully differentiable neuro-soft-symbolic architecture for deductive reasoning over high-dimensional perceptual data.

Other Hacker News

July 23 1985, Commodore introduced its Amiga 1000: 10 years ahead of its time

A retrospective on the 1985 introduction of the Commodore Amiga 1000, highlighting its advanced capabilities for its time.

AI/ML arXiv cs.AI

Long-Term Sequential Decision Making under Risk

Proposes ERQDP, a sampling-free method for long-term sequential decision making under risk using rank-quantile surrogates and Dynamic Programming.

AI/ML arXiv cs.AI

MOF-Sleuth: Tool-Grounded Reward Alignment for Explainable Fine-Grained MOF CIF Auditing

Introduces MOF-Sleuth, an RL-guided auditing agent that uses a forensic lab and reasoning engine to identify errors in crystallographic information files (CIFs).

AI/ML arXiv cs.AI

SenWorld: A Digital-Twin Simulation for Generating Context-Rich Evaluation Data

Presents SenWorld, a digital-twin simulation used to generate context-rich, privacy-safe evaluation data for smartphone personal assistants.