Conferences, Lectures, & Seminars
Events for March
-
NL Seminar - • Enhancing Machine Translation with Large Language Models via Optimizing In Context Examples and Dictionary Based Prompting
Thu, Mar 09, 2023 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Marjan Ghazvininejad, FAIR, Facebook AI Research
Talk Title: Enhancing Machine Translation with Large Language Models via Optimizing In Context Examples and Dictionary Based Prompting
Abstract: REMINDER:
This Talk will be a Live Broadcast Only, It "Will Not" be recorded.
Meeting hosts only admit guests that they know to the Zoom meeting. Hence, you are highly encouraged to use your USC account to sign into Zoom.
If you are an outside visitor, please inform us at nlg DASH seminar DASH host AT isi DOT edu beforehand so we will be aware of your attendance and let you in.
Large language models LLMs have revolutionized natural language processing by demonstrating impressive abilities to perform a wide range of tasks, including machine translation MT. However, the quality and domain of the in-context examples used to prompt these models can significantly impact their performance for specific tasks. In this talk, I will discuss two recent papers that propose to optimize in-context examples and leverage bilingual dictionaries to enhance the quality and controllability of MT with LLMs. First, I will explore the impact of in-context examples on the translation quality of LLMs and highlight the challenges of selecting good examples in both in-domain and out-of-domain settings. Then, I will discuss how we can leverage bilingual dictionaries to provide fine-grained phrase-level control hints in the prompts of LLMs.
Biography: Marjan Ghazvininejad is a senior research scientist at Facebook AI Research. She received her Ph.D. at the University of Southern California on neural creative language generation. Her research interests include text representation, language generation, and machine translation. Her recent research has focused on how to optimize the use of large language models in various applications.
Host: Jon May and Justin Cho
More Info: https://nlg.isi.edu/nl-seminar/
Webcast: https://usc.zoom.us/j/96863901584Location: Information Science Institute (ISI) - Virtual and ISI-Conf Rm#689
WebCast Link: https://usc.zoom.us/j/96863901584
Audiences: Everyone Is Invited
Contact: Pete Zamar
Event Link: https://nlg.isi.edu/nl-seminar/
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
NL Seminar-Designing and Evaluating Language Models for Human Interaction
Thu, Mar 23, 2023 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Mina Lee, Stanford University
Talk Title: Designing and Evaluating Language Models for Human Interaction
Abstract: REMINDER
Meeting hosts only admit guests that they know to the Zoom meeting. Hence, you are highly encouraged to use your USC account to sign into Zoom.
If you are an outside visitor, please inform us at nlg DASH seminar DASH host AT isi DOT edu beforehand so we will be aware of your attendance and let you in.
Despite the recent advancements in language models LMs, most LMs are not optimized for, nor are they evaluated on, real-world usage with human interaction. In this talk, I will present my research on designing and evaluating LMs for human LM interaction. Concretely, I will first describe how we can support human editing needs by enabling any LM to perform text infilling at any position in a document i.e., fill in the blanks. I will then introduce CoAuthor, a platform for capturing human LM interaction in collaborative writing as rich, replayable, keystroke level interaction traces. With the platform, I demonstrate how collecting a large interaction dataset and analyzing the traces provide unique insights into LM capabilities regarding language, ideation, and collaboration. Lastly, I will propose a new framework, HALIE Human AI Language based Interaction Evaluation, that defines the components of interactive systems and evaluation metrics for human LM interaction beyond writing. I will conclude by discussing open challenges and future directions in this field.
Biography: Mina Lee is a final year Ph.D. candidate at Stanford University, advised by Professor Percy Liang. Her research goal is to design and evaluate language models to enhance our productivity and creativity and understand how these models change the way we write. She has built various writing assistants, including an autocomplete system, a contextual thesaurus system, and a creative story writing system, as well as evaluated language models based on their ability to interact with humans and augment human capabilities.
She was named one of MIT Technology Reviews Korean Innovators under 35 in 2022, and her work has been published in top tier venues in natural language processing e.g., ACL and NAACL, machine learning e.g., NeurIPS, and human computer interaction e.g., CHI. Her recent work on human AI collaborative writing received an Honorable Mention Award at CHI 2022 and was featured in various media outlets including The Economist.
Host: Jon May and Justin Cho
More Info: https://nlg.isi.edu/nl-seminar/
Webcast: https://www.youtube.com/watch?v=rfl3_fa8eHQLocation: Information Science Institute (ISI) - Virtual and ISI-Conf Rm#689
WebCast Link: https://www.youtube.com/watch?v=rfl3_fa8eHQ
Audiences: Everyone Is Invited
Contact: Pete Zamar
Event Link: https://nlg.isi.edu/nl-seminar/
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor. -
NL Seminar-Getting AI to Do Things I Can't
Thu, Mar 30, 2023 @ 11:00 AM - 12:00 PM
Information Sciences Institute
Conferences, Lectures, & Seminars
Speaker: Ruiqi Zhong, Cal-Berekely
Talk Title: Getting AI to Do Things I Can't
Series: NL Seminar
Abstract: REMINDER
Meeting hosts only admit guests that they know to the Zoom meeting. Hence, you are highly encouraged to use your USC account to sign into Zoom.
If you are an outside visitor, please inform us at nlg DASH seminar DASH host AT isi DOT edu beforehand so we will be aware of your attendance and let you in.
Is it possible to tame powerful AI systems even when we struggle to determine the ground truth ourselves? In this talk, I will cover two example NLP tasks 1. automatically searching for goal-relevant patterns in large text collections and explaining them to humans in natural language 2. labeling complex SQL programs using non-programmers with the aid of our AI system and achieving accuracy on par with database experts. In both cases, we build tools that help humans scrutinize the AI's behavior with high effectiveness but low effort, bringing new insights that human experts have not anticipated.
Biography: Ruiqi Zhong is a 4th year Ph.D. student advised by Jacob Steinhardt and Dan Klein, working on NLP and AI Alignment.
Ruiqi Zhong attends th Univ. of California Berkeley, he is working on NLP and AI Alignment. His research aims to enable humans to effectively supervise AI systems on tasks where the ground truth is hard to obtain. He reads about epistemology and labor economy in his spare time.
Host: Jon May and Justin Cho
More Info: https://nlg.isi.edu/nl-seminar/
Webcast: https://www.youtube.com/watch?v=dHkYN33TtLMLocation: Information Science Institute (ISI) - Virtual and ISI-Conf Rm#689
WebCast Link: https://www.youtube.com/watch?v=dHkYN33TtLM
Audiences: Everyone Is Invited
Contact: Pete Zamar
Event Link: https://nlg.isi.edu/nl-seminar/
This event is open to all eligible individuals. USC Viterbi operates all of its activities consistent with the University's Notice of Non-Discrimination. Eligibility is not determined based on race, sex, ethnicity, sexual orientation, or any other prohibited factor.
