Alexander Martin

I am a Ph.D. student at Johns Hopkins University, advised by Dr. Ben Van Durme. My research focuses on multimodal retrieval and understanding, specifically advancing efficienct and scalable collaborative multimodal search and reasoning.

The problems I’m interested in are:

Efficient and scalable multimodal retrieval with late interaction models, multi-vector index compression, and rerankers (1,2) over large-scale multilingual video collections that model complex user needs (1,2).
Collaborative article generation in multimodal query-driven article generation (1,2) and interpretable continual document authoring.
Evaluations for text generation from any modality and interpreting RAG evals.

My keywords are: multimodal retrieval, search, video understanding, multimodal RAG, efficiency, collaborative generation.

Before Johns Hopkins, I got my B.S. from the University of Rochester advised by Dr. Jiebo Luo and Dr. Aaron Steven White.

[Resume]

news

Apr 28, 2026	1 paper accepted at EMNLP 2026!
Apr 28, 2026	1 paper accepted at SIGIR 2026!
Apr 27, 2026	1 paper accepted at ACL 2026!
Feb 26, 2025	2 for papers at CVPR 2025 1 2!
Aug 26, 2024	Starting Ph.D. at JHU

selected publications

Knowledge Pull Requests for Continual Document Authoring

Alexander Martin, and Benjamin Van Durme

2026
Seeing Through the MiRAGE: Evaluating Multimodal Retrieval Augmented Generation

Alexander Martin, William Walden, Reno Kriz, Dengjia Zhang, Kate Sanders, Eugene Yang, Chihsheng Jin, and Benjamin Van Durme

2026
Multi-Vector Index Compression in Any Modality

Hanxiang Qin^*, Alexander Martin^*, Rohan Jha, Chunsheng Zuo, Reno Kriz, and Benjamin Van Durme

2026
WikiVideo: Article Generation from Multiple Videos

Alexander Martin, Reno Kriz, William Gantt Walden, Kate Sanders, Hannah Recknor, Eugene Yang, Francis Ferraro, and Benjamin Van Durme

2025

PDF
Video-ColBERT: Contextualized Late Interaction for Text-to-Video Retrieval

Arun Reddy^*, Alexander Martin^*, Eugene Yang, Andrew Yates, Kate Sanders, Kenton Murray, Reno Kriz, Celso M Melo, Benjamin Van Durme, and Rama Chellappa

In IEEE Conference on Computer Vision and Pattern Recognition, Jun 2025

Abs PDF

In this work, we tackle the problem of text-to-video re- trieval (T2VR). Inspired by the success of late interaction techniques in text-document, text-image, and text-video re- trieval, our approach, Video-ColBERT, introduces a simple and efficient mechanism for fine-grained similarity assess- ment between queries and videos. Video-ColBERT is built upon three main components: a fine-grained spatial and temporal token-wise interaction, query and visual expan- sions, and a dual sigmoid loss during training. We find that this interaction and training paradigm leads to strong in- dividual, yet compatible representations for encoding video content. These representations lead to increases in perfor- mance on common text-to-video retrieval benchmarks com- pared to other bi-encoder methods.