Dan R. DeGenaro

Poulton Hall 240

1421 37th Street NW

Washington, DC 20057

I’m an incoming PhD student at Georgetown University, where I will be working with Dr. Sarah Bargal on multimodal intelligent systems – those that integrate text, vision, and other forms of data such as audio. I’m also affiliated with the PICoL Lab led by Dr. Ethan Wilcox.

I am interested in the development of safe, ethical, and energy-efficient multimodal intelligent systems that serve the needs of everyday people while respecting important rights such as privacy, copyright, and the right to be forgotten.

I am also interested in low-resource machine translation and speech recognition, multilingual NLP, and information-theoretic approaches to language modeling and linguistics.

news

Jun 25, 2025	Teaching as a MITES Semester Project Course Instructor for the second year running!
May 16, 2025	Completed my Master of Science in Computational Linguistics!
May 15, 2025	Workshop paper accepted to ACL 2025!
Apr 22, 2025	Won Georgetown University’s Graduate Student Teaching Assistant Award!
Apr 04, 2025	Demo paper accepted to SIGIR 2025!

latest posts

Apr 19, 2025	Principal component analysis of BERT's static embeddings
Mar 26, 2025	Creating a distribution with a specific entropy using PyTorch
Mar 21, 2025	Lecture on n-grams

selected publications

NAACL
Experiments in Mamba Sequence Modeling and NLLB-200 Fine-Tuning for Low Resource Multilingual Machine Translation

Dan DeGenaro, and Tom Lupicki

In Proceedings of the 4th Workshop on Natural Language Processing for Indigenous Languages of the Americas (AmericasNLP 2024), Jun 2024

Abs DOI Bib PDF

This paper presents DC_DMV‘s submission to the AmericasNLP 2024 Shared Task 1: Machine Translation Systems for Indigenous Languages. Our submission consists of two multilingual approaches to building machine translation systems from Spanish to eleven Indigenous languages: fine-tuning the 600M distilled variant of NLLB-200, and an experiment in training from scratch a neural network using the Mamba State Space Modeling architecture. We achieve the best results on the test set for a total of 4 of the language pairs between two checkpoints by fine-tuning NLLB-200, and outperform the baseline score on the test set for 2 languages.
@inproceedings{degenaro-lupicki-2024-experiments, title = {Experiments in Mamba Sequence Modeling and {NLLB}-200 Fine-Tuning for Low Resource Multilingual Machine Translation}, author = {DeGenaro, Dan and Lupicki, Tom}, editor = {Mager, Manuel and Ebrahimi, Abteen and Rijhwani, Shruti and Oncevay, Arturo and Chiruzzo, Luis and Pugh, Robert and von der Wense, Katharina}, booktitle = {Proceedings of the 4th Workshop on Natural Language Processing for Indigenous Languages of the Americas (AmericasNLP 2024)}, month = jun, year = {2024}, address = {Mexico City, Mexico}, publisher = {Association for Computational Linguistics}, url = {https://aclanthology.org/2024.americasnlp-1.22/}, doi = {10.18653/v1/2024.americasnlp-1.22}, pages = {188--194}, }
SIGIR
MMMORRF: Multimodal Multilingual Modularized Reciprocal Rank Fusion

Saron Samuel, Dan DeGenaro, Jimena Guallar-Blasco, and 12 more authors

Jun 2025

Abs DOI Bib PDF

Videos inherently contain multiple modalities, including visual events, text overlays, sounds, and speech, all of which are important for retrieval. However, state-of-the-art multimodal language models like VAST and LanguageBind are built on vision-language models (VLMs), and thus overly prioritize visual signals. Retrieval benchmarks further reinforce this bias by focusing on visual queries and neglecting other modalities. We create a search system MMMORRF that extracts text and features from both visual and audio modalities and integrates them with a novel modality-aware weighted reciprocal rank fusion. MMMORRF is both effective and efficient, demonstrating practicality in searching videos based on users’ information needs instead of visual descriptive queries. We evaluate MMMORRF on MultiVENT 2.0 and TVR, two multimodal benchmarks designed for more targeted information needs, and find that it improves nDCG@20 by 81% over leading multimodal encoders and 37% over single-modality retrieval, demonstrating the value of integrating diverse modalities.
@misc{samuel2025mmmorrfmultimodalmultilingualmodularized, title = {MMMORRF: Multimodal Multilingual Modularized Reciprocal Rank Fusion}, author = {Samuel, Saron and DeGenaro, Dan and Guallar-Blasco, Jimena and Sanders, Kate and Eisape, Oluwaseun and Reddy, Arun and Martin, Alexander and Yates, Andrew and Yang, Eugene and Carpenter, Cameron and Etter, David and Kayi, Efsun and Wiesner, Matthew and Murray, Kenton and Kriz, Reno}, year = {2025}, eprint = {2503.20698}, archiveprefix = {arXiv}, primaryclass = {cs.CV}, url = {https://arxiv.org/abs/2503.20698}, doi = {https://doi.org/10.48550/arXiv.2503.20698}, }