Penn BSE in AI

22 papers from researchers affiliated with the ASSET Center and across Penn have been accepted to this year's NeurIPS conference.

NeurIPS is one of the most prestigious platforms for groundbreaking research in artificial intelligence and machine learning. This year, Penn researchers are presenting work on topics such as deep learning, generative models, AI ethics, medical image analysis, and more.

🔗 See the full list of accepted papers: https://bit.ly/3CwfqrD

🎉 Congratulations to all the researchers involved!

BSE in AI students got a sneak peek at the Amy Gutmann Hall, exploring its innovative design and features. The buzz is growing as the official opening in January gets closer.

Meet Archana Vemulapalli (GEN’01, CGS’05), a Penn Engineering alumna and AI leader at AMD, the high-performance computing company and one of the leading makers of graphics processing units (GPUs), the specialized chips powering today’s AI revolution.

“The potential of AI is tremendous,” says Vemulapalli, whose previous roles at AWS, IBM and the municipal government of Washington, D.C., allowed her to demonstrate the benefits of leveraging AI and machine learning (ML) to accelerate progress in a wide range of fields.

Reflecting on her time at Penn, she says, “Penn gave me a very well-rounded perspective of how to deepen your engineering skills, but also how to frame engineering in the context of what it means for business, which I think has been immensely valuable for me throughout my career.”

Her education at Penn Engineering has played a crucial role in her journey to the forefront of AI. “Penn gave me a very well-rounded perspective of how to deepen your engineering skills,” she says, “but also how to frame engineering in the context of what it means for business, which I think has been immensely valuable for me throughout my career.”

Read her story: https://bit.ly/3AiLPkM


Last week, hosted its annual Family Weekend, featuring a range of events on campus.

Vijay Kumar, Nemirovsky Family Dean, and Robert Ghrist, Andrea Mitchell University Professor and Associate Dean for Undergraduate Education, updated the Engineering School, highlighting the new BSE in AI program.

A student panel also shared insights into their internship experiences, engaging families in meaningful discussions about the opportunities available within the program.

Danaë Metaxa, Raj and Neera Singh Term Assistant Professor in Computer and Information Science (CIS), will represent Penn in the U.S. AI Safety Institute Consortium (AISIC). This consortium unites AI creators, researchers, and civil society to promote the development of safe and trustworthy artificial intelligence.

Metaxa brings their expertise in bias in AI-based sociotechnical systems, such as targeted ads and social media platforms, and will contribute their expertise on algorithm auditing, a method they have explored through sociotechnical auditing research.

Read the full press release: https://bit.ly/4dJVNte

Current AI models in medical imaging often struggle with accuracy. This limitation can be particularly problematic when dealing with diverse patient populations.

To address this challenge, Mark Yatskar, Assistant Professor in Computer and Information Science (CIS) and Chris Callison-Burch, Professor in CIS, along with doctoral student Yue Yang, developed a new approach called Knowledge-enhanced Bottlenecks (KnoBo).

KnoBo requires AI to make decisions based on established medical knowledge, mirroring the thought process and training of human physicians. This approach increases accuracy, in particular when patient race and s*x differ from the training data, and improves the transparency of AI decision making.

With a projected shortage of 80,000 physicians in the U.S. by 2036, there is hope that this innovative model can open the door to the safe application of AI in medicine.

The 2024 Nobel Prize in Chemistry honors groundbreaking advancements in AI, with David Baker recognized for his achievement in designing entirely new proteins, and Demis Hassabis & John M. Jumper awarded for their AI model, AlphaFold2, which solves the 50-year-old problem of predicting protein structures. 🧬

As noted in the Nobel announcement, "Life could not exist without proteins. That we can now predict protein structures and design our own proteins confers the greatest benefit to humankind."

Last Friday, the PRECISE Center hosted its Industry Day, bringing together industry leaders to discuss crucial topics such as building trustworthy AI, applications in banking and healthcare, AI's impacts on the environment, and global regulations and collaborations on AI technology.

Aruna Miller, Lieutenant Governor of Maryland, and Vijay Kumar, Nemirovsky Family Dean of Penn Engineering, delivered opening remarks. The event also featured a keynote address by Dragoș Tudorache, Member of the European Parliament, alongside other distinguished industry leaders. A big thank you to all the speakers and attendees who participated in this important conversation.

John J. Hopfield and Geoffrey E. Hinton have won the 2024 Nobel Prize in Physics for their groundbreaking work in artificial neural networks. Their research laid the foundation for many modern AI applications. Congratulations to these pioneers for their incredible contributions in artificial intelligence!

Can AI transform cardiovascular risk assessment through eye imaging?

A new study reveals that artificial intelligence can analyze retinal images to assess HbA1c levels, a key indicator for diabetes and cardiovascular disease. Lead researcher Lama Al-Aswad, Professor of Ophthalmology at the Scheie Eye Institute, and K*k Jin Jang, a postdoctoral researcher in the PRECISE Center, explore the potential of this technology.

The findings suggest a promising path for early disease detection and personalized healthcare solutions. Follow the link to read more. https://bit.ly/4eQrS3C

Can computers accelerate antibiotic discovery by mining the vast troves of biological data?

César de la Fuente, Presidential Associate Professor at Penn Engineering, is leading a team that uses AI to uncover new antibiotics from ancient molecules. Their tool, Antibiotic Peptide de-Extinction (APEX), employs advanced algorithms to analyze vast amounts of genetic data from extinct organisms like woolly mammoths, dramatically reducing the time required for the preclinical discovery phase of antibiotic research.

With over 7 million deaths from bacterial infections last year—two-thirds of which were due to antimicrobial resistance—innovative approaches in medicine are urgently needed. De la Fuente’s work combines ancient biology with modern technology to tackle these pressing health challenges.

Read the full article in The Philadelphia Citizen: https://bit.ly/3XUus1o