Research Feed Page 9
Search source-linked summaries of recent AI and machine-learning papers by topic, by date, or by whether they include code or a diagram.
Research results
The paper introduces a rank-calibrated detector called TRACE-C designed to identify anomalies in complex electricity system telemetry.
Tydra combines transformer and state-space architectures to achieve faster inference on tabular data than the existing TabPFN foundation model.
The paper introduces a GUI-grounded simulation agent that uses pixel-level perception and reinforcement learning to generate authentic, multi-turn e-commerce shopping trajectories.
The paper introduces a method to improve privacy policy adherence in LLMs by intervening on specific attention heads to align model outputs with user-defined privacy preferences.
EviRank improves image retrieval accuracy by replacing unstructured reasoning with a structured, criteria-based verification framework.
InfinityEdit uses a lightweight adapter to enable consistent, long-term video editing for continuous data streams.
This research evaluates how effectively various language models automate the generation of compliance documentation like digital product passports and data protection assessments.
TLive-Omni is a multimodal model designed to process and understand the complex mix of speech, text, and video signals found in e-commerce live streams.
The COEC method improves the accuracy of pruned large language models by using a specialized, two-sided rotation technique that avoids the pitfalls of direct weight refitting.
The authors introduce a multi-agent system called Patent-MAF that processes raw invention disclosures into formal patent specifications and claims.
The paper introduces E2-TTT, a new method for Test-Time Training that uses chunk-wise updates to achieve higher performance while maintaining computational efficiency.
OmniAssistBench is a new evaluation framework designed to measure how well multimodal AI models handle complex, multi-turn interactions in real-time video scenarios.
The researchers investigated whether software migration specifications are interchangeable across different LLM development agents, finding that they are not agent-neutral artifacts.
AID-Guard ensures that AI agent decisions result in exactly one provider effect by binding user intent to durable, stateful authorization protocols.
The authors introduce a method to compress token data in vision-language-action models by identifying and prioritizing information that has the least impact on physical robot movements.
The paper demonstrates that existing content screening and provenance ranking methods fail to reliably defend agent memory systems from adversarial data injection.
The paper introduces RARE, a method to steer Mixture of Experts models by decoupling control interventions from the model router mechanism to maintain performance and reliability.
EnSI-RAG improves long-document question answering by indexing documents based on structured entity relationships rather than simple text chunks.
The paper introduces an agent that evaluates Retrieval-Augmented Generation outputs by combining document screening and factual verification to block poisoned data and instruction injection.
A single researcher successfully used AI agents to develop a complete system from application code to silicon tapeout in five weeks.