publications & articles | Jaspreet Ranjit

2025

Uncovering Intervention Opportunities for Suicide Prevention with Language Model Assistants

Jaspreet Ranjit, Hyundong J. Cho, Claire J. Smerdon, and 5 more authors

In EAAMO’25, GenAI4Health NeurIPS’25, 2025

Runner up for best doctoral oral presentation @ ShowCAIS’25 Abs Code

Runner up for best doctoral oral presentatino at ShowCAIS 2025

The National Violent Death Reporting System (NVDRS) documents in- formation about suicides in the United States, including free text narra- tives (e.g., circumstances surrounding a suicide). In a demanding public health data pipeline, annotators manually extract structured information from death investigation records following extensive guidelines developed painstakingly by experts. In this work, we facilitate data-driven insights from the NVDRS data to support the development of novel suicide inter- ventions by investigating the value of language models (LMs) as efficient assistants to these (a) data annotators and (b) experts. We find that LM predictions match existing data annotations about 85% of the time across 50 NVDRS variables. In the cases where the LM disagrees with existing an- notations, expert review reveals that LM assistants can surface annotation discrepancies 38% of the time. Finally, we introduce a human-in-the-loop algorithm to assist experts in efficiently building and refining guidelines for annotating new variables by allowing them to focus only on providing feed- back for incorrect LM predictions. We apply our algorithm to a real-world case study for a new variable that characterizes victim interactions with lawyers and demonstrate that it achieves comparable annotation quality with a laborious manual approach. Our findings provide evidence that LMs can serve as effective assistants to public health researchers who handle sensitive data in high-stakes scenarios.

2024

OATH-Frames: Characterizing Online Attitudes Towards Homelessness via LLM Assistants

Jaspreet Ranjit, Brihi Joshi, Rebecca Dorn, and 6 more authors

In Proceedings of EMNLP, 2024

Outstanding Paper Award @ EMNLP 2024; Best Poster @ ShowCAIS’24 Abs Blog Code

Outstanding Paper Award @ EMNLP 2024; Jaspreet received a best poster award at USC CAIS’s annual symposium, ShowCAIS in Spring 2024.

Homelessness in the U.S. is widespread; individual beliefs and attitudes towards homelessness—often expressed on social media are complex and nuanced (e.g. critical as well as sympathetic). Such attitudes can be challenging to summarize at scale, obfuscating the broader public opinion which advocacy organizations use to guide public policy and reform efforts. Our work proposes an approach to enable a large-scale study on homelessness via two major contributions. First, with the help of domain experts in social work and their trainees, we characterize Online Attitudes towards Homelessness in nine hierarchical frames (OATH-Frames) on a collection of 4K social media posts. Further, in an effort to ease the annotation of these frames, we employ GPT-4 as an LLM assistant to the experts; GPT-4 + Expert annotation presents an attractive trade off owing to a 6.5× speedup in annotation time despite only incurring a 2 point F1 difference in annotation performance. Our effort results in a collection of 8K social media posts labeled by domain and trained experts (with and without GPT-4 assistance). Second, using predicted OATH-Frames on a Flan-T5-Large model trained on our data, we perform a large-scale analysis on 2.4M posts on homelessness. We find that posts that contain mentions of west coast states express more harmful generalizations of people experiencing homelessness (PEH) compared to posts about east coast states. We also find marked differences in attitudes across vulnerable populations as they are compared to PEH as being either more or less deserving of aid.

2023

Variation of Gender Biases in Visual Recognition Models Before and After Finetuning

Jaspreet Ranjit, Tianlu Wang, Baishakhi Ray, and 1 more author

Workshop on Algorithmic Fairness through the Lens of Time at NeuRIPS, 2023

Abs HTML

We introduce a framework to measure how biases change before and after fine-tuning a large scale visual recognition model for a downstream task. Deep learning models trained on increasing amounts of data are known to encode societal biases. Many computer vision systems today rely on models typically pretrained on large scale datasets. While bias mitigation techniques have been developed for tuning models for downstream tasks, it is currently unclear what are the effects of biases already encoded in a pretrained model. Our framework incorporates sets of canonical images representing individual and pairs of concepts to highlight changes in biases for an array of off-the-shelf pretrained models across model sizes, dataset sizes, and training objectives. Through our analyses, we find that (1) supervised models trained on datasets such as ImageNet-21k are more likely to retain their pretraining biases regardless of the target dataset compared to self-supervised models. We also find that (2) models finetuned on larger scale datasets are more likely to introduce new biased associations. Our results also suggest that (3) biases can transfer to finetuned models and the finetuning objective and dataset can impact the extent of transferred biases.

2021

Scenario2Vector: Scenario Description Language Based Embeddings for Traffic Situations

Aron Harder, Jaspreet Ranjit, and Madhur Behl

In Proceedings of the ACM/IEEE 12th International Conference on Cyber-Physical Systems, 2021

Abs HTML

A popular metric for measuring progress in autonomous driving has been the "miles per intervention". This is nowhere near a sufficient metric and it does not allow for a fair comparison between the capabilities of two autonomous vehicles (AVs). In this paper we propose Scenario2Vector - a Scenario Description Language (SDL) based embedding for traffic situations that allows us to automatically search for similar traffic situations from large AV data-sets. Our SDL embedding distills a traffic situation experienced by an AV into its canonical components - actors, actions, and the traffic scene. We can then use this embedding to evaluate similarity of different traffic situations in vector space. We have also created a first of its kind, Traffic Scenario Similarity (TSS) dataset which contains human ranking annotations for the similarity between traffic scenarios. Using the TSS data, we compare our SDL embedding -with textual caption based search methods such as Sentence2Vector. We find that Scenario2Vector outperforms Sentence2Vector by 13% ; and is a promising step towards enabling fair comparisons among AVs by inspecting how they perform in similar traffic situations. We hope that Scenario2Vector can have a similar impact to the AV community that Word2Vec/Sent2Vec have had in Natural Language Processing datasets.
Uncovering bias in search and recommendations

Jaspreet Ranjit

Vimeo Engineering Blog on Medium, 2021

HTML

2019

Anchorless object detection for 3D point cloud object detection

Jaspreet Ranjit, and Andrew Draganov

Expedition Technology Blog, 2019

HTML