CMU researchers propose miniCodeProps: a minimal AI benchmark for testing code properties
Recently, ai Agents have shown very promising developments in automating the proving of mathematical theorems and verifying the correctness of ...
Tag: CMU
CMU researchers propose XGrammar: an open source library for efficient, flexible and portable structured generation
The field of structured generation has become important with the rise of LLMs. Capable of generating human-like text, these models ...
MBZUAI and CMU researchers present Bi-Mamba: a scalable and efficient 1-bit Mamba architecture designed for large language models in multiple sizes (780M, 1.3B and 2.7B parameters)
The evolution of machine learning has brought significant advances in language models, which are essential for tasks such as text ...
CMU researchers propose OpenFLAME: a federated and decentralized location service
Maps are widely used today and are useful in numerous location-based applications, including navigation, ride-sharing, fitness tracking, gaming, robotics, and ...
Snowflake and CMU researchers present SuffixDecoding: a new model-free approach to accelerate large language model (LLM) inference using speculative decoding
Large Language Models (LLM) have quickly become a critical component of today's consumer and enterprise applications. However, the need for ...
CMU researchers explore expert guidance and strategic deviations in multi-agent imitation learning
The problem of a mediator learning to coordinate a group of strategic agents is posed through recommendations of actions without ...
CMU researchers propose XEUS: a multilingual universal speech encoder trained on over 4000 languages
Self-supervised learning (SSL) has expanded the reach of speech technologies to many languages by minimizing the need for labeled data. ...
This AI paper from CMU and Google DeepMind studies the role of synthetic data in improving the mathematical reasoning abilities of LLM students
Large language models (LLMs) face a critical challenge in their training process: the looming shortage of high-quality Internet data. Predictions ...
CMU Researchers Introduce TriForce: A Hierarchical Speculative Decoding AI System That Is Scalable to Long Sequence Generation
With the widespread implementation of large language models (LLMs) for long content generation, there is a growing need for efficient ...
CMU researchers present 'Echo Embeddings': an embedding strategy designed to address an architectural limitation of autoregressive models
Neural text embeddings play a critical role in many modern natural language processing (NLP) applications. These embeddings are like fingerprints ...
Recommended.
