American researchers at MIT, Brigham and Women’s Hospital, and Duke University have designed an innovative approach to identifying the transporters used by different drugs that are taken orally. The strategy involves the use of both tissue models and machine-learning algorithms, and has already revealed that a commonly prescribed antibiotic and a blood thinner can interfere…
Large language AI models are notorious for crashing or slowing down during lengthy human-AI dialogues, posing a major barrier to the effective use of chatbots in many applications. Now, a team of researchers from MIT and other institutions propose a novel solution - by modifying the key-value cache, or the 'conversation memory', they improved the…
MIT researchers have developed a new tiny cryptographic ID tag with revolutionary terahertz technology, making it smaller, cheaper, and more secure than traditional radio frequency tags (RFIDs). The latter are often attached to products to verify authenticity but can easily be compromised by counterfeiters who remove and reattach them to fake products. To combat this,…
MIT researchers have discovered a fault in the design of language machine-learning models that can cause AI chatbots' performance to drastically deteriorate during lengthy conversations. Essentially, when data stored in a chatbot's "memory" (known as the key-value cache) exceeds its capacity, the earliest data is removed, sometimes causing the chatbot to malfunction or slow down.…
Researchers at MIT have developed a cryptographic ID tag that is significantly smaller, cheaper, and more secure than traditional radio frequency identification (RFID) tags. The new tag is based on the use of terahertz waves which are smaller and have much higher frequencies than radio waves.
The innovation overcomes a major security flaw common with…
Researchers from MIT and other institutions have developed a method that enables a chatbot to carry on unbroken conversation without crashing or losing performance. This method, named StreamingLLM, involves a tweak to the key-value cache, a form of "conversation memory", that helps AI operate. The team found when the cache became too full, the first…
MIT researchers have created a secure and portable ID tag, utilizing terahertz waves, a technology considered superior compared to traditional RFIDs due to the improved security it offers. The team has bypassed the significant vulnerability of both RFID and terahertz tags, where a counterfeiter can remove the tag from an original item and transfer it…
In 2019 Haofeng "Hector" Xu, having a background in aerospace engineering and presently pursuing a PhD in MIT’s Department of Aeronautics and Astronautics, started learning to fly helicopters. The experience, fraught with risk, led him to consider how helicopter flight could be made safer. In 2021 Xu founded Rotor Technologies, an autonomous helicopter company focused…
Researchers from MIT and other institutions have discovered the key to why AI chatbot conversations can break down and developed a solution that enables continuous dialogue. The issue lies in the chatbot's key-value cache (akin to a conversational memory). In some models, earlier data points are discarded when the cache reaches its limit, causing the…