I'm Helia Hashemi, a Senior Researcher at Microsoft working on Efficient AI. I completed my PhD in Computer Science at the University of Massahcusetts Amherst under the supervision of W. Bruce Croft in the Center for Intelligent Information Retrieval (CIIR). My PhD dissertation is about Retrieval-Augmented Representation Learning for Information Retrieval. During my PhD, I received a Data Science PhD Fellowship from Bloomberg, a Graduate Scholarship from UMass, and a Best Short Paper Honorable Mention award from ICTIR 2019. I completed research internships at Bloomberg AI, Google, Microsoft, and Spotify Research.
In a broad sense, my research interests lie in the areas of Information Retrieval (IR), Natural Language Processing (NLP), and Machine Learning (ML). My work has led to a number of novel retrieval-augmented representation learning approaches for various real-world applications, such as conversational search (SIGIR 2020), web search (CIKM 2021), and text set generation (CIKM 2022). I have also worked on non-factoid question answering (ICTIR 2019, ECIR 2020, ICTIR 2021) and podcast search (KDD 2021, SIGIR 2021).
Email: heliahashemi@microsoft.com
LLM-Rubric: A Multidimensional, Calibrated Approach to Automated Evaluation of Natural Language Text
H. Hashemi*, J. Eisner*, C. Rosset, B. Van Durme, and C. Kedzie.
ACL 2024
Dense Retrieval Adaptation using Target Domain Description
H. Hashemi, Y. Zhuang, S. Kothur, S. Prasad, E. Meij, and W. B. Croft.
ICTIR 2023
Stochastic Optimization of Text Set Generation for Learning Multiple Query Intent Representations
H. Hashemi, H. Zamani, and W. B. Croft.
CIKM 2022
Learning Multiple Intent Representations for Search Queries.
H. Hashemi, H. Zamani, and W. B. Croft.
CIKM 2021
Passage Similarity and Diversification in Non-factoid Question Answering.
L. Vikraman, A. Montazeralghaem, H. Hashemi, W. B. Croft, and J. Allan.
ICTIR 2021
Neural Instant Search for Music and Podcast.
H. Hashemi, A. Pappu, M. Tian, P. Chandar, M. Lalmas, and B. Carterette.
KDD 2021
Current Challenges and Future Directions in Podcast Information Access.
R. Jones, H. Zamani, M. Schedl, C. Chen, S. Reddy, A. Clifton, J. Karlgren, H. Hashemi, A. Pappu, Z. Nazari, L. Yang, O. Semerci, H. Bouchard, and B. Carterette.
SIGIR 2021
Guided Transformer: Leveraging Multiple External Sources for Representation Learning in Conversational Search.
H. Hashemi, H. Zamani, and W. B. Croft.
SIGIR 2020
ANTIQUE: A Non-Factoid Question Answering Benchmark.
H. Hashemi, M. Aliannejadi, H. Zamani, and W. B. Croft.
ECIR 2020