Machine Learning

Researchers from MIT BCS, the University of Cambridge, and the Alan Turing Institute explore the historical pursuit of automated mathematicians in artificial intelligence, emphasizing the recent impact of LLMs. It advocates a cognitive science perspective and...
The global phenomenon of LLM (Large Language Model) products, exemplified by the widespread adoption of ChatGPT, has gathered significant attention. A consensus has emerged among many individuals regarding the advantages of LLMs in comprehending natural language...

This AI Paper Proposes a Novel Pre-Training Strategy Called Privacy-Preserving MAE-Align’ to Effectively Combine Synthetic Data and Human-Removed Real Data

Action recognition, the task of identifying and classifying human actions from video sequences, is a crucial field within computer vision. However, its reliance on...

Google and MIT Researchers Introduce StableRep: Revolutionizing AI Training with Synthetic Imagery for Enhanced Machine Learning

Researchers have explored the potential of using synthetic images generated by text-to-image models to learn visual representations and pave the way for more efficient...

Meet One-2-3-45++: An Innovative Artificial Intelligence Method that Transforms a Single Image into a Detailed 3D Textured Mesh in Approximately One Minute

Researchers from UC San Diego, Zhejiang University, Tsinghua University, UCLA, and Stanford University have introduced One-2-3-45++, an innovative AI method for rapid and high-fidelity...

Meet Slope TransFormer: A Large Language Model (LLM) Trained Specifically to Understand the Language of Banks

In payments, understanding transactions is crucial for assessing risks in businesses. However, deciphering messy bank transaction data poses a challenge, as it is expressed...

Meet LEO: A Groundbreaking Embodied Multi-Modal Agent for Advanced 3D World Interaction and Task Solving

AI systems capable of handling multiple tasks or domains without significant reprogramming or retraining are generalist agents. These agents aim to generalize knowledge and...

Microsoft Researchers Propose PIT (Permutation Invariant Transformation): A Deep Learning Compiler for Dynamic Sparsity

Recently, deep learning has been marked by a surge in research aimed at optimizing models for dynamic sparsity. In this scenario, sparsity patterns only...

McMaster University and FAIR Meta Researchers Propose a Novel Machine Learning Approach by Parameterizing the Electronic Density with a Normalizing Flow Ansatz

Researchers from McMaster University and FAIR Meta have developed a new machine learning (ML) technique for orbital-free density functional theory (OF-DFT). This ML method...

‘Lookahead Decoding’: A Parallel Decoding Algorithm to Accelerate LLM Inference

Although large language models (LLMs) such as GPT-4 and LLaMA are rapidly reimagining modern-day applications, their inference is slow and difficult to optimize because...

ETH Zurich Researchers Introduce UltraFastBERT: A BERT Variant that Uses 0.3% of its Neurons during Inference while Performing on Par with Similar BERT Models

The development of UltraFastBERT by researchers at ETH Zurich addressed the problem of reducing the number of neurons used during inference while maintaining performance...

Unveiling the Frontiers of Scientific Discovery with GPT-4: A Comprehensive Evaluation Across Multiple Disciplines for Large Language Models

Large Language Models (LLMs) have recently gained a lot of appreciation from the Artificial Intelligence (AI) community. These models have remarkable capabilities and excel...

Microsoft Releases Orca 2: Pioneering Advanced Reasoning in Smaller Language Models with Tailored Training Strategies

LLMs (Large Language Models) are trained on vast volumes of textual data to comprehend and produce language similar to that of humans. The GPT-3,...

ByteDance Introduces PixelDance: A Novel Video Generation Approach based on Diffusion Models that Incorporates Image Instructions with Text Instructions

A team of researchers from ByteDance Research introduces PixelDance, a video generation approach that utilizes text and image instructions to create videos with diverse...

Recent articles