Scientists from MIT and the University of Basel in Switzerland have introduced a novel machine-learning framework that employs generative artificial intelligence (AI) models to automatically map out phase diagrams for novel physical systems. This groundbreaking approach addresses the challenge of quantifying phase changes in complex systems with limited data.
Phase transitions, such as the freezing of water, are commonplace, but detecting phase changes in novel materials or intricate physical systems presents unique challenges. Traditional manual techniques rely heavily on theoretical expertise and can be time-consuming. To overcome these limitations, the researchers turned to generative AI models to develop a more efficient and data-driven approach.
Their framework, detailed in a paper published in Physical Review Letters, leverages generative models to recognize phases and detect transitions in physical systems. Unlike conventional machine-learning techniques that require extensive labeled datasets, this approach utilizes physics-informed machine learning and does not depend on large training datasets.
The researchers demonstrated the effectiveness of their method in detecting phase transitions by identifying order parameters that signify changes in the system. By incorporating knowledge about the physical system directly into the machine-learning scheme, the framework outperforms traditional techniques and enhances computational efficiency.
Moreover, this approach opens up possibilities for various binary classification tasks in physical systems, such as detecting entanglement in quantum systems or selecting the most suitable theoretical model for a given problem. It could also contribute to improving large language models like ChatGPT by optimizing parameters for better performance.
Looking ahead, the researchers aim to explore theoretical guarantees regarding the number of measurements required to detect phase transitions effectively and estimate the computational resources needed for implementation.
Funding for this research was provided by the Swiss National Science Foundation, the MIT-Switzerland Lockheed Martin Seed Fund, and MIT International Science and Technology Initiatives.
Source: news.mit.edu

The post Scientists use generative AI to answer complex questions in physics appeared first on HIPTHER Alerts.

The Advanced Research Projects Agency – Infrastructure (ARPA-I) of the United States Department of Transportation (USDOT) is inviting input from interested parties regarding the potential utilization of artificial intelligence (AI) within transportation. They are also seeking insights into the emerging challenges and opportunities associated with the development and implementation of AI technologies across all modes of transportation.
The objective of this Request for Information (RFI) is to gather feedback from a diverse range of stakeholders regarding AI opportunities, challenges, and associated matters in transportation, in accordance with Executive Order (EO) 14110 of October 30, 2023, titled “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.”
Interested parties are encouraged to submit written comments electronically to Docket Number DOT–OST–2024–0049 via the Federal eRulemaking Portal. Comments must be received by July 2, 2024. Submissions, excluding personal information, will be made available to the public on regulations.gov, as per DOT’s Privacy Act Statement.
For inquiries regarding this RFI, individuals may contact [email protected]. Additionally, Mr. Timothy A. Klein, Director of Technology Policy and Outreach at the Office of the Assistant Secretary for Research and Technology, can be reached at 202-366-0075 or via email at [email protected].
Source: traffictechnologytoday.com

The post USDOT seeks input on effective and safe AI use in transportation appeared first on HIPTHER Alerts.

In 2016, James Vlahos faced heartbreaking news – his father received a terminal cancer diagnosis.
“I loved my dad, I was losing my dad,” recalls James, based in Oakland, California.
Determined to cherish the time he had left with his father, James embarked on an oral history project, spending countless hours audio recording his father’s life story. This endeavor coincided with James’ burgeoning interest in AI, prompting him to ponder the possibility of creating something interactive from the recordings.
“I thought, gosh, what if I could make something interactive out of this?” he muses. “For a way to more richly keep his memories, and some sense of his personality, which was so wonderful, to keep that around.”
Although James’ father, John, passed away in 2017, James had transformed the recorded memories into an AI-powered chatbot capable of answering questions about his dad’s life – in his father’s voice.
While the concept of using AI to emulate deceased loved ones has long been explored in science fiction, advancements in AI technology have brought it into reality. In 2019, James launched HereafterAI, allowing users to create similar chatbots for their own departed loved ones.
James acknowledges that while the chatbot doesn’t erase the pain of his father’s death, it provides him with solace and an interactive repository of memories to cherish.
Meanwhile, South Korea’s DeepBrain AI takes this concept further by creating video-based avatars of deceased individuals, capturing their likeness, voice, and mannerisms with striking accuracy.
“We are cloning the person’s likeness to 96.5% of the similarity of the original person,” explains Michael Jung, DeepBrain’s chief financial officer. “So mostly the family don’t feel uncomfortable talking with the deceased family member, even though it is an AI avatar.”
DeepBrain envisions its technology as part of a “well dying” culture, where individuals prepare for death in advance, leaving behind a living legacy of family histories and memories.
However, this technology comes at a significant cost, with users paying up to $50,000 (£39,000) for the filming process and avatar creation. Despite the steep price tag, investors remain bullish on its potential, evident in DeepBrain’s substantial fundraising success, having raised $44m in its last funding round.
Source: bbc.com

The post The man who turned his dead father into a chatbot appeared first on HIPTHER Alerts.