Researchers from the Massachusetts Institute of Technology (MIT) have developed an antitampering ID tag that uses terahertz waves to make products almost impossible to counterfeit. The tag uses microscopic metal particles mixed with the glue that attaches the tag to a product. The terahertz waves can detect the unique pattern created by the metal particles,…
Researchers from MIT, Brigham and Women’s Hospital, and Duke University have developed a strategy to identify the transporter proteins used by different drugs to exit the digestive tract. This knowledge could improve patient treatment, as drugs relying on the same transporter can interfere with each other and should not be prescribed together.
The researchers adapted…
A team of researchers, including those from the Massachusetts Institute of Technology (MIT), have created a system called StreamingLLM that allows chatbots to maintain ongoing dialogues without suffering from performance issues. The method involves a reconfiguration of the model's key-value cache—a form of memory storage—that commonly leads to models failing when the cache is overloaded…
MIT researchers have developed an antitampering cryptographic ID tag that offers improved security, lower cost and a much smaller size than traditional radio frequency tags. The previous generation of terahertz tags, like radio frequency identifiers (RFIDs), were vulnerable to tampering, as fraudsters could move the tags from genuine to counterfeit goods without the authentication system…
A team of researchers from MIT, Brigham and Women’s Hospital, and Duke University have worked together to develop a method for identifying how different drugs exit the digestive system. This is a key question in pharmacology, as it is often associated with drug interactions and efficacy. In this study, the researchers utilized a combination of…
Researchers from MIT and other institutions have proposed a solution to the challenge of AI systems losing the context of conversation in extended dialogues. Large language models such as ChatGPT, which enable the functioning of chatbots, often struggle to retain information from long conversations, resulting in rapid performance deterioration.
The team has developed a method…
Researchers at MIT have developed a new cryptographic ID tag that is cheaper, smaller, and more secure than traditional radio frequency tags (RFIDs). Initially, the researchers developed a tiny tag using terahertz waves, which offered improved security over RFIDs. However, this tag shared a vulnerability with RFID tags: a counterfeiter could peel off the tag…
Scientists from MIT, Brigham and Women’s Hospital, and Duke University have developed a method to track how oral drugs pass through the human digestive tract. This research is considered key as it can help predict potential drug interactions which improve patient treatment. The research used a combination of tissue models and machine learning algorithms.
Oral…
A group of researchers from MIT and other institutions have pinpointed a key issue that causes performance degradation in AI chatbots during long conversations and have developed a simple solution to rectify it. Large language machine-learning models such as the ChatGPT use key-value cache to store data. However, when the cache needs to hold more…
Researchers at MIT have developed a highly advanced anti-tampering ID tag that is significantly smaller and cheaper than traditional Radio Frequency Identification (RFID) tags. It leverages the power of terahertz waves to improve upon conventional security tools and offers an innovative solution to safeguard items from counterfeit.
Traditional security tags, much like this one, use radio…
Researchers from MIT, Brigham and Women’s Hospital, and Duke University have developed a research approach to identify how different drugs exit the digestive tract. The method uses tissue models and machine-learning algorithms to understand which transporters are used by drugs, revealing how a commonly prescribed antibiotic and blood thinners can interfere with each other. Transporter…
Researchers from MIT and other institutions have devised an innovative solution to prevent chatbots from crashing during prolonged dialogues. The method, known as StreamingLLM, makes a simple adjustment to the key-value cache, essentially the 'conversation memory,' of well-developed machine-learning models. By ensuring the first few data points don't get bumped out, the chatbot can maintain…