Agrima Seth

Agrima Seth

Ph.D. student, School of Information

Talks and presentations

Your spouse needs professional help: Determining the contextual appropriateness of messages through modeling social relationships

August 07, 2023

Talk, Cohere for AI, Virtual

In this talk, I present the paper published at ACL 2023, in which we introduce a new approach to identifying inappropriate communication by explicitly modeling the social relationships between individuals. We introduce a new dataset of contextually situated judgments of appropriateness and show that large language models can readily incorporate relationship information to identify appropriateness in a given context accurately. LINK

Anomaly Detection using Machine Learning for Data Quality Monitoring in the CMS Experiment

August 17, 2017

Talk, CERN openlab summer students' lightning talks 1, Geneva, Switzerland

This talk is about our project that aimed at applying recent progress in Machine Learning techniques (unsupervised machine learning method - autoencoder) to the automation of quality assessment. In our project, we concentrated on analyzing the occupancy of the drift tube chambers. The aim was to check large volumes of data in real-time and improve the ability to detect unexpected features.