The Archive

Benchmark of 6 deep learning weather emulators (Pangu, FuXi, GraphCast, Aurora) on extreme heat forecasting at 10-15 day lead times.

Inverted causal self-attention mechanism for discovering causal structures in high-dimensional multivariate time series.

Vibe-FDTR agent framework enables LLMs to automate reproducible thermal property analysis from natural language requests.

Compressed LLMs pass quality guards but hallucinate procedure steps in agentic execution, exposing fidelity-safety gap.

Privacy-preserving federated learning framework for clinical EEG using masking-based secure aggregation and threshold secret sharing.

LLM pipeline for automating MADRS depression assessment scoring in clinical trial structured interviews.

Black-box evaluation reveals commercial multimodal content moderation APIs vulnerable to simple image transformations.

Gaussian noise injection prevents oversmoothing in deep recurrent GNNs by preserving representation diversity.

ReAlloc framework uses causal inference for multi-channel marketing budget allocation in e-commerce platforms.

TPACK-guided empirical study examines multi-agent AI tool integration into software engineering requirements quality curriculum.

AgenticASR refines speech transcription via agent-based approach that revises text when context clarifies from later speech.

Unified algorithmic framework for optimal decision trees compares search strategies to improve scalability.

Candidate-aware decoding enables adaptive early-exit in diffusion language models by position-specific commitment decisions.

The Series A was led by Khosla Ventures, with participation from Allianz, Rebel Fund, Brick and Mortar Ventures’ Darren Bechtel, and Y Combinator’s Garry Tan.

RRM augments multimodal memory graphs with reflection mechanisms for long-horizon video reasoning agent adaptation.

OPLD uses on-policy latent distillation to train flexible latent visual reasoning without external trace supervision.

Ablation study on deep learning architectures for automated TCM tongue diagnosis achieves F1 0.66 on 5k-image dataset.

ParliamentBench evaluates deceptive reasoning in 16 LLMs via Secret Hitler game framework with 1,600 adversarial matches.

Philosophical critique argues LLM anthropomorphisms (hallucination, agency, sentience) reflect category mistakes in asymmetric human-machine language games.

LM-GRASP reformulates combinatorial optimization as online imitation learning, training Transformer constructive policies from scratch per problem instance.

Pre-registered audit compares LLM tutoring pedagogical vs. direct-answer policies; Claude Opus 4.8 and GPT-5.6 Sol judge helpfulness signals.

FinSMART applies market-aligned reinforcement learning to financial sentiment analysis for algorithmic trading.

ConMem framework estimates diagnostic value of inspection logs for LLM-assisted steel-equipment screening and early-risk detection.

Label-free criterion selects optimal unsupervised domain adaptation algorithm and hyperparameters for medical imaging without target-domain labels.

Information bottleneck framework ensures faithful time-series forecast explanations via interpretable-by-design approach with faithfulness guarantees.

Comparative study finds LLMs and linguist trainees struggle similarly with subjective Appraisal theory annotation in TED talk discourse.

AI engineers adopt ontologies to constrain probabilistic agents within deterministic logical boundaries, reviving semantic web techniques.

It is impossible to make large language models fully secure against hacks because of a fundamental flaw in how they work, a team of researchers argue in a paper presented at the International Conference on Machine Learning, a top AI conference, this month. The claim has huge implications for the safety of this technology, which…

OpenAI reduces GPT-5.6 pricing for Luna and Terra variants, emphasizing efficiency gains for enterprise AI deployment.

Microsoft pitched its own homegrown AI models, harnesses, and even a Mythos competitor on Wednesday, telling Wall Street it plans for continued growth.