Latest in AI and developer tools

Today's trail

updated 27m ago

Last hour Today

Pulse · rebellionresearch.com

rebellionresearch.com · Nov 1

How to Secure AI Agents: Identifying Hidden Risks in Third-Party AI Skills

Pulse · techtimes.com

techtimes.com · Nov 1

Parse and Nomic Launch Cascade: AI Immune Drug Discovery Gets Its Missing Protein Signal

AI drug discovery single-cell datasets now have their first systematic secretome training layer: Parse Biosciences and Nomic Bio launched Cascade, pairing 100 million RNA-seq profiles with matched 1,000-protein secretome reads across 150,000-plus immune perturbation conditions, backed by GSK,

Pulse · techgig.com

techgig.com · Nov 1

Spring ecosystem releases first release candidates for 17 projects

The Spring ecosystem announced the first release candidates for 17 projects, including Spring Boot 4.2.0-RC1 and Spring Framework 7.1.0-RC1. These updates deliver bug fixes, dependency upgrades, and new features across various components. Developers can expect improved performance and enhanced functionalities in these latest releases.

Pulse · msn.com

msn.com · Oct 31

Taiwan's MediaTek eyes 20% share of global AI chip market

MediaTek (TWSE: 2454) expects to capture about 20% of the global market for application-specific integrated circuits (ASICs) used in AI accelerators next year, Focus Taiwan reported on October 31. Chief Executive Officer Rick Tsai told investors the custom chip should bring in about $16bn in sales in 2027,

Pulse

Top picks this weekarXiv

arXiv

Oct 29

Elucidating the Design Space of LM Data Attribution

Data attribution methods for Language Models (LMs) are typically evaluated using downstream tasks like fact retrieval. However, existing evaluations vary multiple dataset properties simultaneously, making it impossible to isolate exactly why an attribution algorithm fails. To diagnose these failures, we explore the design space of LLM data attribution, with a primary focus on evaluation. We introduce the Controlled Attribution Benchmark (CATT) for fact tracing, in which we fine-tune LLMs on entirely fictitious knowledge, providing unambiguous ground truth for attribution. This synthetic setup allows us to systematically alter one dataset configuration at a time, such as document redundancy and lexical distractors. With this configurable evaluation, we demonstrate that redundancy and distractors consistently degrade attribution performance. Rather than proposing a new estimator, we examine two design choices that are underexplored: the loss used to compute gradients and how per-token attribution scores are aggregated into a document score. We find that the true causal evidence within a training document is highly localized, but standard methods dilute this evidence into noise by…

Top picks

All arXiv Hugging Face GitHub Hacker News

Latest Top this week

arXiv

Sep 20

Perplexity Predicts Protection: Choosing Pretrained Backbones for Worst-Client Fairness in Federated Parameter-Efficient Fine-Tuning

Federated learning lets multiple parties train a shared model without pooling their data, but a client with far less data than the others can end up poorly served even when the group's average accuracy looks fine. We ask whether the choice of pretrained backbone affects this…

arXiv

Sep 20

Long-Tail Rebalancing for Non-Verbal Vocalization-Aware ASR: A Track 1 System for the NVVSpeech Challenge

Non-verbal vocalizations (NVVs) carry important paralinguistic information but are often omitted by conventional automatic speech recognition (ASR) systems. The ISCSLP NVVSpeech Challenge requires joint transcription of lexical content and 16 NVV categories under limited and…

arXiv

Sep 20

TRACE: Tractable Routing Autoencoder for Clinical ECG

Deep learning has advanced automated electrocardiogram (ECG) diagnosis, but the field's most accurate models, foundation models pretrained on millions of recordings, are not decision-pathway auditable: a clinician cannot trace a diagnosis to a physiological pathway or intervene…

arXiv

Sep 20

RLVR$^{2}$: Reinforcement Learning with Verifiable Rubric-based Ranking

Reinforcement Learning with Verifiable Rewards (RLVR) is expanding from tasks with well-defined correctness signals, such as mathematics and code, toward multifaceted quality requirements specified by multi-dimensional rubrics. Since policy optimization consumes one scalar per…

This week in Multimodal

This week

Get the morning digest

Today's trail

Pulse

Top picks

Today's trail

This week in Multimodal

Most active this week

Rising tools