Category
425 papers
Results indicate that current LLMs provide uneven safety assurance across Urdu's script varieties, with smaller open-weight models showing substantially higher instability and missed-harm rates than frontier closed models.
This work analyzes a complete corpus of 10,211 inbound scam and spam calls collected over 54 days by an AI voice-agent honeypot that answered callers and kept them talking, and introduced in a companion data descriptor.
Constrained-guided mapping is proposed, a neuro-symbolic method with three stages: schema-grounded admissibility constraints with metadata mc =, where tau_c denotes the constraint type and delta_c provides executable relation and normalization logic, and constraint-restricted candidate generation with cascade relaxation to guarantee a nonempty feasible set under noise.
Findings support a hybrid paradigm in which AI augments, but does not replace, health economists in value assessment and formulary decision support within managed care settings.
Reach audiences
Advertise in front of researchers, engineers, and readers.
A hybrid model, termed SE_LLM_ST, is introduced, which leverages recent advancements in large language models (LLMs) and transformer-based architectures to effectively capture both contextual and sequential information vital for Persian emotion recognition.
MEPO-SLM is presented, a framework that reformulates prompt engineering for SLMs as a four-objective Pareto problem over task inaccuracy, and Phi-3-mini and Gemma-2B on English TriviaQA and Arabic medical QA, and TinyLlama-1.1B on TriviaQA only are evaluated.
A role-specialized Mixture-of-Agents (MoA) that combines medical knowledge retrieval with contrastive similar-patient reasoning is studied, placing role design as a key factor in privacy-constrained, training-free clinical LLM prediction.
Mathematics has often been organized around an authorial subject: one person, or a small group, composing proofs through language, notation, and judgment. Large language models, proof assistants, formal libraries, and repositories now make another production unit technically credible: a human-machine assemblage. This article calls that unit a studio ecobiont and asks when it is epistemically legitimate. Its governance thesis is that human participation is substantive only when the system preserves traceable provenance, reconstructible human competence, capacity to challenge the result, effective authority to stop or withdraw it, and public responsibility. These conditions distinguish a governed studio from a degenerate studio whose human oversight is ceremonial. A comparison of Polymath, the Liquid Tensor Experiment, Danus, and the Jacobian counterexample episode shows that collaboration, formalization, technical orchestration, and epistemic governance are independent dimensions. The proposed understanding audit and contribution-authority trace are governance designs, not validated measures. No causal superiority over authorial practice is claimed.
SchemaGUI, a template-based benchmark for controllable GUI generation evaluation, synthesizing paired natural language instructions and deterministic function-call references from parameterized interface schemas can generate thousands of deterministically annotated tasks in seconds without human labeling.
This work introduces Instruction-Followed Function Calling (IFFC), a novel framework that decouples function-calling logic from the primary LLM and delegates it to a dedicated smaller model operating within the instruction-following paradigm, establishing a new paradigm for reliable, resource-efficient function calling in edge-computing scenarios.
Dual-Layer Agentic Memory is proposed, a framework that shifts memory management to the write phase through cost-aware epistemic routing and periodic parametric consolidation, allowing the router to adaptively suppress redundant writes as the model's epistemic boundaries evolve.
Two state-of-the-art multimodal models, Gemma-3 and Qwen-VL, are assessed on their ability to interpret mechanical problem images by eliciting a step-by-step chain of thought (CoT) and a final answer, and final answers are compared to verified solutions to measure accuracy.
Data Management
What if pathology foundation models could do more with less? GigaPath-Flash and GigaTIME-Flash cut computational demands while maintaining strong performance, opening the door to larger studies and broader exploration. The post GigaPath-Flash and GigaTIME-Flash: Toward population-scale discovery with efficient pathology foundation models appeared first on Microsoft Research.
With millions of users across the world, Julia has been used to conduct cutting-edge research and to design new drugs, jet engines, heat pumps, and more.
Earth AI
Health & Bioscience