Sarah Alnegheimish

I'm a Ph.D. candidate at MIT pursing degree in Electrical Engineering and Computer Science (EECS), advised by Dr. Kalyan Veeramachaneni, and a member of the Data to AI Lab at LIDS. My research interest is in the field of applied deep learning, particularly within time series data. Recently, I have been investigating the efficacy of foundation models for unsupervised time series anomaly detection. I am also interested in open-source development and in building systems that make machine learning usable.

Selected Projects

Large Language Models for Time Series
Repurposing LLMs for time series tasks. SigLLM is a framework to probe LLMs in finding anomalies in zero-shot. It includes a time-series-to-text conversion module, as well as two paradigms of detection: prompt engineering and forecasting. [paper] [code] [press]
Unsupervised Time Series Anomaly Detection
Developing Orion (1,000 github stars), a machine learning library for unsupervised time series anomaly detection. Given time series data, users select from an array of pipelines, including TadGAN and AER, to identify rare patterns. [paper] [tadgan paper] [aer paper] [code] [blog] [press]
Biases in Large Language Models
Analyzing and evaluating gender-profession bias in language models by curating a dataset of real-world natural prompts. Comparing template– and natural-based prompts that elicit biased output. [paper] [code]

Selected Publications

Making the End-User a Priority in Benchmarking: OrionBench for Unsupervised Time Series Anomaly Detection
Sarah Alnegheimish, Laure Berti-Equille, Kalyan Veeramachaneni 2024 IEEE International Conference on Big Data.
Can Large Language Models be Anomaly Detectors for Time Series?
Sarah Alnegheimish, Linh Nguyen, Laure Berti-Equille, Kalyan Veeramachaneni 2024 IEEE International Conference on Data Science and Advanced Analytics (DSAA 2024).
Using Natural Sentence Prompts for Understanding Biases in Language Models
Sarah Alnegheimish, Alicia Guo, Yi Sun 2022 Conference of the North American Chapter of the Association for Computational Linguistics (NAACL).
Sintel: A Machine Learning Framework to Extract Insights from Signals
Sarah Alnegheimish, Dongyu Liu, Carles Sala, Laure Berti-Equille, Kalyan Veeramachaneni 2022 ACM SIGMOD International Conference on Management of Data.