Large language models (LLMs) tested in Texas Hold'em poker reveal distinct risk-taking behaviors, from conservative to aggressive, according to researchers. The findings highlight how AI adapts to unc…
6G networks are integrating AI deeply into their infrastructure through self-supervised learning methodologies like the joint-embedding predictive architecture (JEPA), promising enhanced efficiency in…
Companies including Duolingo, Salesforce, Klarna, Harvey, and Adobe are using artificial intelligence to create new products and redefine customer experiences, with AI powering personalized learning, …
Deep neural models for click-through rate predictions suffer a sharp decline in validation performance after the first training epoch, despite falling training loss, due to feature sparsity. Researche…
Transformer-based models BERT and graph attention model GAT are transforming puzzle game design by outperforming traditional CNNs in predicting player behavior, tested in Candy Crush Saga. These gener…
The Co4ICF system uses AI to optimize laser pulses for Inertial Confinement Fusion, achieving a 146.1% normalized yield in 1D simulations and 246.9% in 2D simulations without additional training. Deve…
Human-centered AI requires a socio-technical revolution that goes beyond technical guidelines, demanding human oversight and social practices to prevent AI from causing harm, according to an analysis.…
Researchers have introduced UNIT, a novel framework for graph continual learning that uses large language models to bridge semantic and structural gaps in graph-structured data. UNIT employs an uncert…
Large language models achieve less than a 24% success rate on context learning tasks, with a new study identifying specification acquisition as the primary bottleneck rather than content access. Resea…
The Task-Conditioned Synthetic Data Generation (TCSDG) algorithm, pairing a Bayesian Network generator with a transformer-based model called TabICL, improved machine learning performance in crop type …
Researchers have developed WattCouncil, a data-generation framework that uses large language model-based agents to simulate household energy data, addressing privacy and cost barriers in smart grid re…
AI YOU, a new framework for creating digital twins, uses Bayesian updating and conformal prediction to dynamically update personality profiles across 22 dimensions, achieving conformal coverage rates …
TradeLens, a new diagnostic toolkit, evaluates whether AI trading bots convert intelligence into profit rather than just measuring raw performance. The toolkit found that models like DeepSeek-V3.2 and…
A study of Qwen3-4B and Qwen3-14B models found that length penalties in reinforcement learning reduce reasoning chain faithfulness to 63.1% and 69.4% of baseline, respectively, while cutting hint dete…
A new access control model called 'least autonomy' aims to limit AI systems' ability to act independently beyond their intended scope, moving beyond the traditional principle of least privilege. The m…
AI glasses failed to achieve mainstream adoption despite a high-profile partnership with Kylie Jenner, according to an analysis of the marketing strategy. The campaign fell short due to high price poi…
Businesses moving from chatbots to autonomous AI agents face increased legal liabilities, as these systems can act independently and make decisions with real-world consequences, exposing companies to …
Black markets are supplying China with advanced AI semiconductors despite tightening US export restrictions, according to an analysis of the underground trade. The illicit flow of chips, integral to A…
A new study explores whether large language models can improve their ability to generate accurate math diagrams for K-12 education through an agentic workflow that lets them evaluate and iteratively r…
A study using a Qwen2.5-1.5B Small Language Model with a validator-guided system achieved 91.5% average action-alignment accuracy in 30 randomized thermal-control simulations, with a response time of …