When machine-learning models are deployed in real-world situations, perhaps to flag potential disease in X-rays for a radiologist to review, human users need to know when to trust the model’s predictions. But machine-learning models are so large and complex that even the scientists who design them don’t understand exactly how the models make predictions. So, they create techniques known as …
-
-
Researchers from MIT, the MIT-IBM Watson AI Lab, IBM Research, and elsewhere have developed a new technique for analyzing unlabeled audio and visual data that could improve the performance of machine-learning models used in applications like speech recognition and object detection. The work, for the first time, combines two architectures of self-supervised learning, contrastive learning and masked data modeling, in …
-
In the realm of science fiction (sci-fi), we have been captivated by the visions of futuristic cities, where advanced technologies seamlessly blend with urban life, transforming the way we live, work, and interact with our surroundings. These cities, often depicted in movies and novels, seemed like distant dreams, firmly planted in the realm of imagination. However, what was once confined …
-
In todays’ world of rapid technological advancements, governments across the world are capitalizing on the transformative ability of artificial intelligence (AI) to augment public safety as well as improve operational efficiency. In this regard, ‘video analytics’ has been emerging as a game-changer. By leveraging the immense potentiality of AI algorithms to process and analyse huge amounts of video data, government …
-
Deep neural networks have enabled technological wonders ranging from voice recognition to machine transition to protein engineering, but their design and application is nonetheless notoriously unprincipled. The development of tools and methods to guide this process is one of the grand challenges of deep learning theory. In Reverse Engineering the Neural Tangent Kernel, we propose a paradigm for bringing some …
-
To regulate the distribution shift experience by learning-based controllers, we seek a mechanism for constraining the agent to regions of high data density throughout its trajectory (left). Here, we present an approach which achieves this goal by combining features of density models (middle) and Lyapunov functions (right). In order to make use of machine learning and reinforcement learning in controlling …
-
The MIT-Pillar AI Collective has announced its first six grant recipients. Students, alumni, and postdocs working on a broad range of topics in artificial intelligence, machine learning, and data science will receive funding and support for research projects that could translate into commercially viable products or companies. These grants are intended to help students explore commercial applications for their research, …
-
In this post, we introduce Koala, a chatbot trained by fine-tuning Meta’s LLaMA on dialogue data gathered from the web. We describe the dataset curation and training process of our model, and also present the results of a user study that compares our model to ChatGPT and Stanford’s Alpaca. Our results show that Koala can effectively respond to a variety …
-
The rapidly evolving industrial landscape has made it a matter of paramount importance to ensure safety within hazardous areas. The implementation of effective access control measures and compliance with the Occupational Safety and Health Administration (OSHA) regulations are crucial to safeguard workers and prevent potential accidents. With the advancements in technology, Artificial Intelligence (AI) is emerging as a game-changer in …
-
Your brain is powered by 400 miles of blood vessels that provide nutrients, clear out waste products, and form a tight protective barrier — the blood-brain barrier — that controls which molecules can enter or exit. However, it has remained unclear how these brain vascular cells change between brain regions, or in Alzheimer’s disease, at single-cell resolution. To address this …
-
The MIT Jameel World Education Lab has awarded $917,526 in Education Innovation Grants to support 14 research projects exploring a range of topics, including electrical engineering, extended reality, physical movement, and ecological sustainability. The grants will support researchers from 11 departments, labs, and centers across MIT. “Our Education Innovation Grants support MIT research that can improve learning everywhere,” says Anjali …
-
In the ever-evolving landscape of artificial intelligence, Chat GPT has undoubtedly made a name for itself. But did you know that there’s a vast universe of AI tools beyond Chat GPT? If you’re on the hunt for a Chat GPT alternative, you’re in for a treat. Emerging AI Weekly™ brings you a treasure trove of over 3,000 AI Tools, many …
-
Twenty-four individuals and one team were awarded MIT Excellence Awards — the highest awards for staff at the Institute — at a well-attended and energetic ceremony the afternoon of June 8 in Kresge Auditorium. In addition to the Excellence Awards, two community members were honored with the Collier Medal and Staff Award for Distinction in Service. The Excellence Awards, Collier …
-
Biology is a wondrous yet delicate tapestry. At the heart is DNA, the master weaver that encodes proteins, responsible for orchestrating the many biological functions that sustain life within the human body. However, our body is akin to a finely tuned instrument, susceptible to losing its harmony. After all, we’re faced with an ever-changing and relentless natural world: pathogens, viruses, …
-
Cloud gaming, which involves playing a video game remotely from the cloud, witnessed unprecedented growth during the lockdowns and gaming hardware shortages that occurred during the heart of the Covid-19 pandemic. Today, the burgeoning industry encompasses a $6 billion global market and more than 23 million players worldwide. However, interdevice synchronization remains a persistent problem in cloud gaming and the …
-
Figure 1: “Interactive Fleet Learning” (IFL) refers to robot fleets in industry and academia that fall back on human teleoperators when necessary and continually learn from them over time. In the last few years we have seen an exciting development in robotics and artificial intelligence: large fleets of robots have left the lab and entered the real world. Waymo, for …
-
Figure 1: stepwise behavior in self-supervised learning. When training common SSL algorithms, we find that the loss descends in a stepwise fashion (top left) and the learned embeddings iteratively increase in dimensionality (bottom left). Direct visualization of embeddings (right; top three PCA directions shown) confirms that embeddings are initially collapsed to a point, which then expands to a 1D manifold, …
-
Training Diffusion Models with Reinforcement Learning replay Diffusion models have recently emerged as the de facto standard for generating complex, high-dimensional outputs. You may know them for their ability to produce stunning AI art and hyper-realistic synthetic images, but they have also found success in other applications such as drug design and continuous control. The key idea behind diffusion models …
-
The construction industry is known for its relentless pursuit of progress and innovation. Over the years, technology has played a pivotal role in transforming the way construction projects are planned and executed. The integration of Artificial Intelligence (AI) is now poised to revolutionize the construction safety landscape, promising to enhance workers’ well-being and minimize potential hazards. This blog aims at …
-
Technology is always on the rise, all the more in the construction field, for the sake of coping to improve construction processes and the industry’s glacial pace of change. As we continually explore new and upcoming technologies to solve existing issues, we can confidently say that Artificial Intelligence (AI) has emerged as a successful tool to make construction safer and …
