Latest Articles for Performance

Momentum-SAM offers an efficient alternative to traditional training methods for neural networks.

Meta prompting improves language model performance through task breakdown and expert collaboration.

This framework enhances how knowledge is combined in machine learning models for better performance.

This article investigates methods to predict the effectiveness of personalized search results.

Learn how to improve transformer model efficiency with GPU-friendly design.

This study examines how language structure boosts layout predictions in machines.

Combining models to improve predictions of disease outbreaks.

This paper examines the effectiveness of AI-generated text versus human-written content.

A new dataset for testing RAG systems on complex multi-hop questions.

New models aim to improve language tech for Brazilian Portuguese speakers.

A new system optimizes language models for faster, more efficient responses.

This study explores generalizing learned constraints in ASP for better performance in dynamic problems.

Learn how ACADL simplifies modeling for AI hardware accelerators.

Introducing Z3alpha, a new method for SMT strategy generation using Monte Carlo Tree Search.

Using data analysis to identify promising startups before investment.

Improving model accuracy for rare categories in long-tailed datasets.

This research focuses on fair predictive modeling considering hierarchical group structures.

Analyzing GPT-NeoX and LLaMA models for materials science applications.

This research examines how robot foot design affects walking efficiency on sand.

A method for speeding up large language models without sacrificing output quality.

New framework enhances data selection and graph cleaning in noisy environments.

A strategy to enhance performance and fairness in federated learning models.

A method for verifying machine learning models to enhance trust and transparency.

A new framework learns effective starting points for faster fixed-point optimization.

This article presents a benchmark to assess large language models with complex tasks.

This article presents strategies and a new method for effective sample selection in few-shot learning.

This approach enhances efficiency in matroid optimization using quick but imprecise oracles.

A new system enhances performance and reduces delays in medical AI applications.

A project aims to automate train delay code classification using machine learning.

SEABO generates rewards from expert data, simplifying offline imitation learning.

CURE improves robot performance by optimizing configuration settings efficiently.

Cannikin improves DNN training efficiency in diverse machine environments.

Discover how Actions World Models improve decision-making in complex environments.

New methods help robots explore unfamiliar environments without prior maps.

Soft prompts enhance speech recognition technology for better performance in noisy environments.

Longer instructions enhance language model performance and reduce complexity.

New strategies enhance consumer management for improved message processing.

A new method enhances control of nonlinear systems with changing parameters.

This article examines the gap between generative and evaluative abilities of AI models.

Study questions the need for optimizing in-context examples when clear instructions are provided.