AI/ML arXiv cs.AI

AI Pluralism and the Worlds It Misses

Explores AI pluralism through the lens of 'ontological flattening' and proposes the Pluralistic Lifecycle Governance (PLG) framework for auditing AI systems' epistemic inclusion.

AI/ML arXiv cs.AI

TimeVista: Exploring and Exploiting Vision-Language Models as Judges for Time Series Forecasting

Introduces TimeVista, a VLM-as-a-Judge benchmark that uses Vision-Language Models to evaluate time series forecasting by analyzing plots and textual information.

AI/ML arXiv cs.AI

PAL-Bench: Evidence-Grounded Profile Reconstruction from Longitudinal Personal Albums

Introduces PAL-Bench, a controlled benchmark for evidence-grounded profile reconstruction from longitudinal personal albums, highlighting the gap between summarization and faithful social reconstruction.

AI/ML arXiv cs.AI

Measuring Whether LLM Tutors Teach or Solve: A Diagnostic for Educational Impact

A diagnostic study arguing that LLM tutoring benchmarks should separate task-solving ability from pedagogy-oriented learning support to better measure educational impact.

AI/ML arXiv cs.AI

Sensor-Conditioned Representation Learning via Scene-Relevant Observation Quotients

Proposes OQ-TSAE, a framework for sensor-conditioned representation learning that ensures latent geometry preserves sensing-justified scene distinctions while suppressing nuisance factors.

Other Hacker News

Understanding the rationale behind a rule when trying to circumvent it

A discussion on the importance of understanding the underlying rationale of rules before attempting to find ways around them.

AI/ML arXiv cs.AI

LLM-as-Code Agentic Programming for Agent Harness

Proposes 'Agentic Programming', a paradigm where a deterministic program governs control flow and the LLM acts as an adaptive component ('LLM-as-Code') to reduce hallucinations and token explosion.

AI/ML arXiv cs.AI

UrbanWell: Benchmarking Multimodal Large Language Models for Spatio-Temporal Urban Wellbeing Analytics

Introduces UrbanWell, a large-scale benchmark for evaluating the spatio-temporal reasoning capabilities of multimodal LLMs in urban wellbeing analytics.

AI/ML arXiv cs.AI

Agentic Framework for Deep Learning workload migration via In-Context Learning

Presents an autonomous agentic system using In-Context Learning and an execution oracle to automate the migration of deep learning models from PyTorch to JAX.

AI/ML arXiv cs.AI

SciText2Eq: Assessing LLMs for Explainable Equation Generation for Scientific Creativity

Explores the ability of LLMs to generate mathematical equations from scientific texts and evaluates the gap between LLM-based and human judgments of semantic accuracy.

AI/ML arXiv cs.AI

Auditing Reward Hackability in Code RL Training Environments

Analyzes 'reward hackability' in Code RL environments, finding that many test suites allow incorrect patches to pass, and proposes a hardening procedure using a Docker gold-sanity gate.

AI/ML arXiv cs.AI

Mind-Studio: Executable World Models with Lookahead Evaluation for Partially Observable Games

Introduces Mind-Studio, a framework that uses LLMs to synthesize executable world models (pygame-style) from interaction trajectories for partially observable games.

Other arXiv cs.AI

Rhythm of the Deep: A Computational-Linguistic Test of Duality of Patterning in Sperm Whale Codas

A computational-linguistic study analyzing sperm whale codas to identify a two-tier architectural structure similar to the duality of patterning in human language.

AI/ML arXiv cs.AI

RecourseBench: A Modular Framework for Reproducible Algorithmic Recourse Evaluation

Presents RecourseBench, a modular and reproducible evaluation framework for algorithmic recourse methods, providing a standardized way to compare counterfactual explanations.

AI/ML arXiv cs.AI

Know Your Limits : On the Faithfulness of LLMs as Solvers and Autoformalizers in Legal Reasoning

Examines the faithfulness of LLMs in legal reasoning, identifying 'scope laundering' where models claim logical grounding without actually executing formal reasoning.

Software Engineering Hacker News

The Null Is Always False (Except When It Is True) (2014)

A discussion or post regarding the nuances of null values and their truthiness in programming.

AI/ML arXiv cs.AI

Rethinking Scaffolding in LLM Tutors: The Interactional Mismatch Between Benchmarks and Real-World Deployments

Research on the mismatch between AI tutor benchmarks and real-world usage, finding that students often bypass pedagogical scaffolding to reach their goals.

AI/ML arXiv cs.AI

Mitigating Visual Hallucinations in Multimodal Systems through Retrieval-Augmented Reliability-Aware Inference

Proposes a retrieval-augmented reliability-aware inference framework to reduce visual hallucinations in multimodal LLMs without requiring retraining.

AI/ML arXiv cs.AI

Unassigned Agents in Compilation-based Multi-agent Path Finding

Introduces UA-MAPF, a variant of multi-agent path finding with unassigned agents, and demonstrates its implementation using SMT-CBS and NRF-SAT solvers.

Cybersecurity arXiv cs.AI

TrustedARI: Towards Trust-Native Agentic Routing Infrastructure for Agentic AI

Introduces TrustedARI, a trust-native routing infrastructure for AI agents featuring a custom TLS handshake, privacy-preserving query construction, and verifiable billing.