<html><head><meta http-equiv="content-type" content="text/html; charset=utf-8"></head><body dir="auto"><div dir="ltr"></div><div dir="ltr">The seminar will actually be in GHC 4405 today! Sorry for the confusion.</div><div dir="ltr"><br></div><div dir="ltr"><blockquote type="cite">On Feb 12, 2023, at 12:10 PM, Asher Trockman <ashert@cs.cmu.edu> wrote:<br><br></blockquote></div><blockquote type="cite"><div dir="ltr"><div dir="ltr">Dear all,<div><br></div><div>Please join us for this Valentine's Day (<b>Tuesday, 2/14</b>) installment of the<b> CMU AI Seminar Series</b> to learn about some research we love.</div><div><div>The seminar will be held in <u>GHC 6115</u> from <b><font color="#ff0000">1</font></b><font color="#ff0000"><b>2:00-1:00 PM (U.S. Eastern time)</b></font> <b>with pizza provided </b>and will<b> </b>be streamed on Zoom. It is sponsored by <a href="https://sambanova.ai/" target="_blank">SambaNova Systems</a>.</div><div><br></div><div>To learn more about the seminar series or to see the future schedule, please visit the <a href="http://www.cs.cmu.edu/~aiseminar/" target="_blank">seminar website</a>.</div><div><br></div><font color="#0b5394"><span style="background-color:rgb(255,255,0)">This Tuesday (2/14), <u>Uri Alon</u></span><b style="background-color:rgb(255,255,0)"> </b><span style="background-color:rgb(255,255,0)">(CMU LTI) will be giving a talk titled </span><b style="background-color:rgb(255,255,0)">"</b><span style="background-color:rgb(255,255,0)"><b>Natural Language Reasoning with Language Models of Code</b></span><b style="background-color:rgb(255,255,0)">".</b></font></div><div><font color="#0b5394"><span style="background-color:rgb(255,255,0)"><br></span><b>Title</b>: Natural Language Reasoning with Language Models of Code</font><div><font color="#0b5394"><br></font><div><font color="#0b5394"><b>Talk Abstract</b>: In this talk, I will show that LMs that were pretrained on *code* can be better natural language reasoners than LMs that were trained (mostly) on natural language, even when the task does not involve source code at all. <br>In a class of structured NL reasoning tasks, I will show how we can frame the task as code generation; this makes LMs of code such as Codex better reasoners than LMs of natural language such as T5 and GPT-3. <br>Another class of mathematical reasoning tasks was recently unlocked by methods that require LLMs to generate their explicit reasoning steps, such as “chain-of-thought” (Wei et al., 2022).<br>Such methods employ LMs for both understanding the problem description by decomposing it into steps, as well as solving each step of the problem. While LMs seem adept at the step-by-step decomposition part, they often make logical and arithmetic mistakes in the solution part. I will show how LMs of *code* can decompose the natural language problem into runnable steps, which allows us to offload the solution to a programmatic runtime such as a Python interpreter. That is, instead of learning to solve the problem directly, we teach the model to generate a program that solves the problem. Across a variety of benchmarks, this approach leads to more accurate results than much larger models such as PALM-540B using chain-of-thought.<br></font></div><div><div><div><font color="#0b5394"><br></font></div><div><font color="#0b5394"><b>Speaker Bio:</b> Uri Alon is a Postdoctoral Researcher at LTI, working with Prof. Graham Neubig on NLP and learning from source code. Previously, he obtained his PhD at the Technion (Israel), where he worked on modeling programming languages and graphs. Currently, he is also interested in the synergy of neural models with symbolic components such as retrieval, programs, and automata. His personal website is at <a href="https://urialon.ml">https://urialon.ml</a>. Feel free to reach out with any questions or comments about the talk. </font><br></div><div><br></div><div><font color="#0b5394"><b>In person:<span class="gmail-Apple-converted-space"> </span></b>GHC 6115</font></div><div><font color="#0b5394"><b>Zoom Link</b>: <a href="https://cmu.zoom.us/j/99510233317?pwd=ZGx4aExNZ1FNaGY4SHI3Qlh0YjNWUT09" target="_blank">https://cmu.zoom.us/j/99510233317?pwd=ZGx4aExNZ1FNaGY4SHI3Qlh0YjNWUT09</a></font></div></div></div></div></div><div><br></div><div>Thanks,</div><div>Asher Trockman</div></div>
</div></blockquote></body></html>