Omri Suissa

Profile

I'm a researcher in Natural Language Processing and its role in multimodality models (e.g., vision), focusing specifically on embedding models for text and images, both individually and collectively (multimodal embeddings). At Brown University, I am working with Prof. Shekhar Pradhan at the Conversational AI Lab on linguistic supervision of multimodal embeddings. During my PhD at Bar Ilan University under Prof. Maayan Zhitomirsky-Geffet and Dr. Avshalom Elmalech, I engaged in applied research such as developing question-answering systems for digital humanities texts, identifying bias in the U.S. patent system using machine learning models, and implementing post-OCR corrections for low-resource languages. My work often involves interdisciplinary collaboration, particularly with cultural heritage experts, to preserve and analyze historical artifacts using AI, as well as fundamental computer science research on embedding models. These papers won several awards and grants, including the DSI Outstanding Paper Award, the President's and Dean's Publication Award (4 times), and the Data Science Initiative Grant for Deep Learning Research.

My research interests focus on the untapped potential in multimodal encoders (i.e., embeddings) with modality supervision. By introducing better numerical representations of the world, these encoders can serve both as inputs to generative models and as standalone lightweight models for low-resource environments (both in compute and data) and for classification tasks. My research agenda is to enrich textual embeddings with world knowledge from different modalities (e.g., images, video, audio, and sensory data) to better align with human perception. This research agenda will develop methods and models with the potential to significantly advance fields such as human-computer interaction (HCI), robotics, LLM reasoning, information retrieval, medical diagnosis, assistive communication, surveillance, cultural studies, and numerous other downstream applications.

I've served as a lecturer in workshops and courses, including Introduction to Efficient LLMs (Brown University), Large Language Models and Generative AI (Brown University), Introduction to Machine Learning (Bar Ilan University), Advanced Python (Bar Ilan University), and Semantic Web (Bar Ilan University).

I have over twenty years of successful industry experience as a computer and data scientist. I am dedicated to driving innovations that are both theoretically sound and practically impactful in these fields. I serve as the VP of R&D at ClearMash (an AI-based company). My extensive hands-on experience allows me to bridge the gap between theoretical concepts and their practical applications, enriching both academic research and industry solutions.

Publications

Around the GLOBE: Numerical Aggregation Question-answering on Heterogeneous Genealogical Knowledge Graphs with Deep Neural Networks

Around the GLOBE: Numerical Aggregation Question-answering on Heterogeneous Genealogical Knowledge Graphs with Deep Neural Networks

Omri Suissa, M. Zhitomirsky-Geffet, Avshalom Elmalech

ACM Journal on Computing and Cultural Heritage 2023

Question answering with deep neural networks for semi-structured heterogeneous genealogical knowledge graphs

Question answering with deep neural networks for semi-structured heterogeneous genealogical knowledge graphs

Omri Suissa, M. Zhitomirsky-Geffet, Avshalom Elmalech

Semantic Web 2022

Text analysis using deep neural networks in digital humanities and information science

Text analysis using deep neural networks in digital humanities and information science

Omri Suissa, Avshalom Elmalech, M. Zhitomirsky-Geffet

J. Assoc. Inf. Sci. Technol. 2021

Optimizing the Neural Network Training for OCR Error Correction of Historical Hebrew Texts

Optimizing the Neural Network Training for OCR Error Correction of Historical Hebrew Texts

Omri Suissa, Avshalom Elmalech, M. Zhitomirsky-Geffet

arXiv.org 2023

Toward a period-specific optimized neural network for OCR error correction of historical Hebrew texts

Toward a period-specific optimized neural network for OCR error correction of historical Hebrew texts

Omri Suissa, M. Zhitomirsky-Geffet, Avshalom Elmalech

ACM Journal on Computing and Cultural Heritage 2022

Gendered Words and Grant Rates: A Textual Analysis of Disparate Outcomes in the Patent System

Gendered Words and Grant Rates: A Textual Analysis of Disparate Outcomes in the Patent System

Deborah Gerhardt, Miriam Marcowitz-Bitton, W. M. Schuster, Avshalom Elmalech, Omri Suissa, Moshe Mash

Toward the optimized crowdsourcing strategy for OCR post-correction

Toward the optimized crowdsourcing strategy for OCR post-correction

Omri Suissa, Avshalom Elmalech, M. Zhitomirsky-Geffet

Aslib Journal of Information Management 2019

AI‐Based Research Tool for Large Genealogical Corpora: The Case of Jewish Communities Worldwide

AI‐Based Research Tool for Large Genealogical Corpora: The Case of Jewish Communities Worldwide

M. Zhitomirsky-Geffet, Omri Suissa