Featured Students

AI Month, NSF Fellowship, & Enhancing Language Models

In celebration of AI Month, we are thrilled to extend our congratulations to Andrew Zhu, under the guidance of Chris Callison-Burch, for his remarkable achievement in receiving the prestigious 2024 NSF Graduate Research Fellowship Program (GRFP). Andrew’s dedication and innovative vision have earned him this esteemed recognition, marking a significant milestone in his academic journey.

Andrew’s research endeavors are poised to address a critical challenge in the realm of Language Model Machines or Large Language Models (LLMs), where despite their formidable capabilities, they often grapple with the issue of ‘forgetting’ when confronted with complex inquiries or tasks spanning extensive timelines. For instance, while humans effortlessly navigate through multi-step queries, LLMs encounter hurdles in retaining the original context across successive stages of information retrieval. Inspired by this gap, Andrew’s previous work, FanOutQA, shed light on the limitations of current models, showcasing the vast barrier between human accuracy and machine performance.

High-level diagram showing the settings in the FanOutQA benchmark: the LLM (robot) must answer a complex question by breaking it down (bottom). There are 3 settings we test LLMs in: closed-book (the LLM cannot use any search tools, only what it was trained on), open-book (the LLM may search Wikipedia), and evidence-provided (the LLM is given the Wikipedia pages containing the answers and must extract the correct answers).

Central to Andrew’s proposed solution is the concept of recursive delegation, a groundbreaking approach aimed at endowing LLMs with the ability to allocate intricate tasks to subordinate LLMs. This hierarchical structure allows for the decomposition of complex problems into more manageable segments, with each ‘child’ LLM equipped to handle its designated subtask. Through this cascading delegation, Andrew envisions a synergistic collaboration among LLMs, comparable to a relay race where each participant contributes to achieving the final objective.

In laying the groundwork for his research, Andrew has developed Kani, a versatile framework designed to facilitate seamless interaction between LLMs and human-written code. This innovative tool empowers researchers to experiment with different LLM architectures while interfacing with Python functions effortlessly. By democratizing access to such tools, Andrew aims to catalyze progress in the field and foster a collaborative ecosystem of exploration and discovery.

Screenshot of an early proof-of-concept recursive delegation system I developed this year. In this example, it’s helping make travel plans for a Japan trip and looking at Shinkansen (bullet train) travel times on the internet. Each node (in the top right) is an independent LLM. (green=in progress, yellow=waiting, gray=completed, blue=selected)
The same as above, but with the code needed to actually use Kani. It’s designed to be easy to pick up while allowing experienced developers to customize the library extensively.
An overview of what the Kani system is comprised of. It manages the chat history and provides it and user-written functions to underlying LLMs (engines) with a common interface.

Andrew’s advisor, Chris Callison-Burch, effectively summarizes the significance of his work, stating,

“Andrew Zhu’s research is at the cutting edge of programming languages, natural language processing, and artificial intelligence. His groundbreaking open-source software package, Kani, elegantly solves a major weakness of large language models (LLMs) like ChatGPT by enabling them to write code and call functions. Kani seamlessly integrates LLMs into Python programs, allowing developers to write functions that can be called by the language model. This game-changing approach combines the strengths of AI and traditional programming, enabling LLMs to interweave natural language generation with complex computations.

Andrew’s work has the potential to revolutionize how we develop AI systems, paving the way for more powerful and versatile AI applications that go beyond ChatGPT’s current abilities of understanding and generating natural language to also include the abilities to create and manipulate code.”

Chris Callison-Burch

As Andrew embarks on this ambitious journey, his work holds the promise of transforming the landscape of artificial intelligence, ushering in a new era of collaboration and problem-solving. We eagerly anticipate the insights and advancements that will emerge from his pioneering efforts, underscoring the transformative potential of AI in shaping our collective future. Once again, congratulations to Andrew Zhu on this well-deserved recognition, and here’s to a future marked by innovation, collaboration, and boundless possibilities in the realm of AI.


[1] Andrew Zhu, Alyssa Hwang, Liam Dugan, and Chris Callison-Burch. 2024. FanOutQA: Multi-Hop, Multi-Document Question Answering for Large Language Models. Arxiv preprint, in submission at ACL 2024.

[2] Andrew Zhu, Liam Dugan, Alyssa Hwang, and Chris Callison-Burch. 2023. Kani: A Lightweight and Highly Hackable Framework for Building Language Model Applications. In Proceedings of the 3rd Workshop for Natural Language Processing Open Source Software (NLP-OSS 2023), pages 65–77, Singapore. Association for Computational Linguistics.


Penn Students Awarded 2024 NSF GRFP

Four students from the University of Pennsylvania have been awarded the prestigious NSF Graduate Research Fellowship Program (GRFP) for 2024. Three of these awardees, Lena Armstrong, Emmanuel Suarez Acevedo, and Stephan Xie, will be heading to other universities for their PhDs, while Edward Zhang will continue his doctoral studies at Penn. Here are their achievements and reflections on this milestone.

Lena Armstrong (HCI)

Lena Armstrong, who works with Danaë Metaxa, will begin her PhD in Computer Science at Harvard this fall. Reflecting on her achievement, Armstrong said, “I feel excited and honored to have received the NSF Graduate Research Fellowship. This award allows me greater access to resources and opportunities to further my research, advocacy, and teaching goals. My aim is to change the narrative of who gets to shape our digital future and create more inclusive experiences in computer science.”

Armstrong’s research focuses on algorithmic fairness, particularly in employment and housing. She has co-authored several papers, including “Navigating a Black Box: Students’ Experiences and Perceptions of Automated Hiring” and “The Silicon Ceiling: Auditing GPT’s Race and Gender Bias in Hiring.”

Danaë Metaxa praised Armstrong, stating, “Lena is a star, one of the most motivated and creative undergraduates I have worked with. Her departure is bittersweet, but I look forward to seeing all she will accomplish at Harvard.”

Emmanuel Suarez Acevedo (Formal Methods, Verification, and Programming Languages)

Emmanuel Suarez Acevedo, advised by Stephanie Weirich, is heading to Cornell for his PhD. On receiving the fellowship, Acevedo remarked, “I was excited and grateful for the support from fellow students and professors at Penn. The GRFP will allow me to focus on my dissertation topic without the worry of finding funding, enabling me to continue working towards broadening participation in computer science.”

Acevedo recently worked on an extension to an intermediate language for static tracking of program behaviors, collaborating with students from Penn and the University of Michigan. This work aims to improve compiler optimizations.

Stephanie Weirich commented, “Emma is heading to Cornell’s PhD program next year and will intern with Kuen-Bang Hou Favonia at the University of Minnesota this summer. He has worked with me on several projects, including enhancing the Call-by-Push-Value type system and mechanizing proofs in Agda.”

Stephan Xie (Machine Learning)

Stephan Xie, who has been working with Aaron Roth, will pursue his PhD in Machine Learning at Carnegie Mellon University. Xie shared his excitement, saying, “Receiving the GRFP was a huge honor. It gives me more confidence in my research direction and opens up many possibilities as I begin my PhD.”

Xie’s research focuses on advancing machine learning decision-making and discovery in real-world applications. He has worked on projects exploring unbiased predictions and improving the reliability of deep learning.

A project I worked on last year with my advisor Prof. Aaron Roth and graduate students Georgy Noarov and Ramya Ramalingam was along this vein, exploring how we could make “unbiased predictions” that lead to strong decision-making guarantees, i.e. low swap regret for all decision-makers. This also led to a number of applications in online combinatorial optimization, extensive-form games, and uncertainty quantification.

Aaron Roth expressed his enthusiasm for Xie’s work, stating, “It was great fun working with Stephan. Together with my PhD students, we wrote what turned out to be my favorite paper in some time. I’m looking forward to following his progress at CMU.”

Edward Zhang (Computer Vision)

Edward Zhang, who will continue his PhD at Penn, expressed his gratitude for receiving the fellowship. “I am incredibly grateful and honored to have received this opportunity. I hope to leverage these funds to take charge of my own research direction,” Zhang said.

Zhang plans to use the funding to investigate the use of visual computing to solve real-world problems in disciplines like medicine, public health, and public policy. He aims to deploy these solutions in economically sustainable and scalable ways.

Live Mas 2024 Renewal Submission – Edward Zhang
A recent project: Ego-Exo4D – Here is an early stage data collection photo. I credit this project for helping me develop my practical knowledge about computer vision and research in general.

Congratulations to all the 2024 NSF GRFP awardees! This is a well-deserved achievement, and we look forward to seeing the remarkable contributions you will make during your PhDs!


Penn HCI & NLP Students Awarded NSF GRFP

Four PhD students from Penn’s Human-Computer Interaction (HCI) and Natural Language Processing (NLP) groups have been awarded the prestigious NSF Graduate Research Fellowship Program (GRFP) for 2024. These awards recognize and support outstanding graduate students in STEM disciplines. Three of the awardees, Ro Encarnación, Hita Kambhamettu, and Princess Sampson, are pioneering research in HCI, while a fourth student, Andrew Zhu, has made significant strides in NLP.

During AI Month, the CIS Blog highlighted Andrew’s research and accomplishments in a previous article. For more details on Zhu’s awards, click here!

Here are their achievements and reflections on this milestone.

Ro Encarnación

Advised by Danaë Metaxa, Ro Encarnación has been focusing on community-engaged public interest technology policy and design. “I took a big risk leaving the workforce to pursue the kind of work I’m interested in. Receiving the award provided validation that I’m on the right track,” said Encarnación. They are currently working on a project under review on emergent auditing behavior in social media.

Metaxa praises Encarnación’s work, noting, “Ro’s recent work has studied how users interrogate the algorithmic systems they interact with in their everyday lives. Her awarded GRFP proposed to build systems for high schoolers and their families to navigate the high-stakes and unequal setting of high school selection. Ro’s experience as a TechCongress fellow positions her to produce impactful research.”

Hita Kambhamettu

Under the guidance of Andrew Head, Hita Kambhamettu is advancing research at the confluence of human-AI interaction and biomedical informatics. Reflecting on the fellowship, she shared, “When I verified that I had indeed received the fellowship, I was overcome with a profound sense of excitement and immense gratitude. This achievement is due to the steadfast support from my advisor, collaborators, peers, and family.”

Kambhamettu plans to develop intelligent systems that help patients make informed health decisions and create explainable machine learning methods for medical research. She recently completed a project introducing “traceable text” to enhance the understanding of AI-generated medical summaries. Head commended her contributions, saying, “Hita is designing human-AI interfaces to serve as just-in-time translators of health records. With the NSF fellowship, she will build on her strong research foundation to re-envision patients’ relationships with their health information.”

Princess Sampson

Princess Sampson, also advised by Danaë Metaxa, expressed their honor in receiving the NSF-GRFP alongside her peers. She highlighted the collective effort, “We formed a weekly check-in group, broke the application into distinct deliverables, and developed a timeline to hold each other accountable. One win would have been a win for all of us, but it’s a joy and a privilege to win together.”

Sampson aims to empower users, especially from marginalized communities, in their interactions with digital ad targeting systems. They are currently working on a system for collaborative, value-driven ad-blocking. Metaxa lauded her dedication, stating, “Princess has been working on tools for empowering users in the context of digital ad targeting systems. Her paper published in FAccT last year was very well-received, and I look forward to seeing the rest of her PhD develop.”

Advisor Reflections

Danaë Metaxa expressed immense pride in the students, saying, “All of us in the Penn HCI group are incredibly proud of our NSF GRFP awardees this year. Ro, Princess, and Hita are all founding members of the HCI group, our first cohort of students; it’s inspiring and affirming to see their success.”

Andrew Head shared similar sentiments about Kambhamettu’s achievements and future potential, “Hita is making great inroads in her research agenda. With the NSF fellowship, she will continue to advance her work and positively impact the field of biomedical informatics.”

These students’ successes highlight the collaborative and supportive environment at Penn HCI and NLP, paving the way for innovative research with real-world impact.


Recognizing Joe Devietti: Faculty Advising Award Recipient

Joseph Devietti, serving as Associate Professor and Undergraduate Curriculum Chair in the CIS Department, has recently been honored with the Ford Motor Company Award for Faculty Advising. This accolade underscores his unwavering commitment to guiding students towards achieving their academic, professional, and personal aspirations.

Specializing in computer architecture and programming languages, particularly in streamlining multiprocessor programming through advancements in computer architectures and parallel programming models, Devietti brings a wealth of expertise to his role.

Having obtained a BSE in computer science and a BA in English from the University of Pennsylvania in 2006, Devietti furthered his education with MS and PhD degrees in computer science and engineering from the University of Washington in 2009 and 2012, respectively.

Devietti’s dedication extends beyond the classroom, as evidenced by his investment in mentoring student projects. His willingness to provide access to his lab’s resources for academic exploration and projects, coupled with his ability to break down and explain the complicated parts of a technical topic, has garnered admiration from students.

Each year, the Penn Engineering undergraduate student body meticulously selects recipients of these awards, acknowledging individuals for their commitment to teaching, mentorship, and advocacy on behalf of students.

Reflecting on his achievement,

“I felt honored to join so many colleagues from CIS who have been previous winners. I’ll try to continue to live up to the high standards we’ve set for advising!”

Joe Devietti

Congratulations to Joe Devietti on this well-deserved honor!