AI

  Researchers from the Intelligent Autonomous Systems Group, Locomotion Laboratory, German Research Center for AI, Centre for Cognitive Science, and Hessian.AI introduced a benchmark to advance research in Imitation Learning (IL) for locomotion, addressing the limitations of existing measures that often focus on simplified tasks. This new benchmark comprises diverse environments, including quadrupeds, bipeds, and…

  In an era dominated by AI advancements, distinguishing between human and machine-generated content, especially in scientific publications, has become increasingly pressing. This paper addresses this concern head-on, proposing a robust solution to identify and differentiate between human and AI-generated writing accurately for chemistry papers. Current AI text detectors, including the latest OpenAI classifier and…

  As technology continues to advance at an astonishing pace, Cerebras Systems and G42 have just taken a giant leap forward in the world of artificial intelligence. In a groundbreaking partnership, they have successfully completed a 4-Exaflop AI supercomputer, marking a significant milestone in the quest for unprecedented computational power. This achievement also signifies the…

  Modern self-driving systems frequently use Large-scale manually annotated datasets to train object detectors to recognize the traffic participants in the picture. Auto-labeling methods that automatically produce sensor data labels have recently gained more attention. Auto-labeling may provide far bigger datasets at a fraction of the expense of human annotation if its computational cost is…

  In deep learning, Transformer neural networks have garnered significant attention for their effectiveness in various domains, especially in natural language processing and emerging applications like computer vision, robotics, and autonomous driving. However, while enhancing performance, the ever-increasing scale of these models brings about a substantial rise in compute cost and inference latency. The fundamental…

  Researchers from MIT, CarperAI, and Parametrix.AI introduced Neural MMO 2.0, a massively multi-agent environment for reinforcement learning research, emphasizing a versatile task system enabling users to define diverse objectives and reward signals. The key enhancement involves challenging researchers to train agents capable of generalizing to unseen tasks, maps, and opponents. Version 2.0 is a…

  The GPT-4 was tested using a public Turing test on the internet by a group of researchers from UCSD. The best performing GPT-4 prompt was successful in 41% of games, which was better than the baselines given by ELIZA (27%), GPT-3.5 (14%), and random chance (63%), but it still needs to be quite there….

  In a groundbreaking development, engineering company Boston Dynamics has integrated ChatGPT, a sophisticated language model developed by OpenAI, into one of its remarkable robots, Spot. This canine-like companion is now equipped to offer guided tours around a building, providing insightful commentary on each exhibit along the way. Spot has undergone a remarkable transformation, now…

  Consistency models represent a category of generative models designed to generate high-quality data in a single step without relying on adversarial training. These models attain optimal sample quality by learning from pre-trained diffusion models and utilizing metrics like LPIPS (learning Perceptual Image Patch Similarity). The quality of consistency models is limited to the pre-trained…

  In Reinforcement learning (RL), effectively integrating human feedback into learning processes has risen to the forefront as a significant challenge. This challenge becomes particularly pronounced in Reward Learning from Human Feedback (RLHF), especially when dealing with multiple teachers. The complexities surrounding the selection of teachers in RLHF systems have led researchers to introduce the…