Mark Crowley | Lab Spotlight : Publications by Josh Sun

PhD candidate Josh Sun has been in the lab for a few years now, first completing his Master’s with us and is now building up his research portfolio for his PhD in a significant way, we have multiple announcements from his research this term so I thought I’d put them all in one post.

First, Josh successfully completed his PhD Proposal Exam unanimously with Category 1 and very supportive feedback from the committee towards the next steps for his thesis research on Vision Language Navigation. Congratulations Josh!

This research already includes full journal paper published in a top IEEE robotics journal (Sun et al., 2026), which Josh will present in the journal track of the prestigious IROS conference in the fall.

Sun, J. Q., Weng, H., Xing, X., Yeum, C. M., & Crowley, M. (2026). View Invariant Learning for Vision-Language Navigation in Continuous Environments. IEEE Robotics and Automation Letters, 11(5), 5861–5868. https://doi.org/10.1109/LRA.2026.3669785

He was doing this all while taking a graduate course and continuing to work on a Mitacs internship which he started in the fall. This internship is for an existing research project I have on use of ML for antibody discovery with a local research company Kisogi Biotechnology.

Josh has been very well received by the company, helping them build improved, practical Data Science tools. He also worked with the company scientists and engineers to access their data for a more academic pursuit, experimenting with novel Preference Model based Deep Learning algorithms. This resulted in a deeper research collaboration with the company for my lab and produced a solidly performing model.

This work has just recently been accepted to the International Machine Learning Conference (ICML) and Josh will present the results at ICML in South Korea over the summer (Sun et al., 2026).

Sun, J. Q., Babaie, M., Hou, W., Crowley, M., & Young, D. (2026). Preference-based Antibody Expression Ranking: Scaling with Large-scale Weak Supervision. Forty-Third International Conference on Machine Learning (ICML), 8. https://openreview.net/forum?id=mrZiCCb3zv

He also completed the final steps of a long publishing process for a paper to the Remote Sensing Journal (Chen et al., 2026). This project applied machine learning methodology to a soil moisture problem in a novel way.

Chen, Y., Tong, C., Sun, J. Q., Shangguan, Y., Deng, X., Crowley, M., Wang, H., Ye, Y., Bao, H., & Huang, R. (2026). A novel soil moisture retrieval method via combining radiative transfer model and machine learning. Remote Sensing of Environment, 338, 115378. https://doi.org/https://doi.org/10.1016/j.rse.2026.115378

In all of these projects, Josh has been the leading force for collaboration, publishing initiative and has done the bulk of the core technical work and writing, with guidance and editing from myself and his collaborators.

References:

A novel soil moisture retrieval method via combining radiative transfer model and machine learning

Yurun Chen, Cheng Tong, Josh Qixuan Sun, Yulin Shangguan, Xiaodong Deng, Mark Crowley, Hongquan Wang, Yang Ye, Haijun Bao, and Ruqi Huang.

Remote Sensing of Environment. 338, 2026.

Abs PDF URL

Soil moisture (SM) is a key variable in the global water cycle, and passive microwave remote sensing is widely used for large-scale SM estimation. Although radiative transfer models (RTM) and machine learning (ML) approaches have advanced SM retrieval, interpretable methods that provide a direct, inspectable mapping from satellite observations to SM remain scarce. In this study, we present an initial attempt to develop such an interpretable retrieval framework by fusing the physical principles of the RTM with the mathematical formulation capability of the Kolmogorov–Arnold Network (KAN). The framework proceeds in two stages: first, a physically consistent training dataset is generated by the RTM that links SM with brightness temperature (TB), surface temperature (ST), and vegetation optical depth (VOD), thereby preserving key physical dependencies; second, KAN is trained on this dataset to extract explicit mathematical expressions that relate TB, ST and VOD to SM for different land-cover types. These derived expressions are applied to SMAP observations to produce global daily SM estimates for 2015–2023. Validation against in situ soil moisture measurements from global networks, including the International Soil Moisture Network (ISMN) and the Qinghai Lake Basin dense network (QLB-NET), shows that the KAN retrieval achieves an average correlation coefficient R = 0.64 and unbiased RMSE (ubRMSE) = 0.07 m3/m3, comparable to the SMAP Level-3 product (R = 0.65, ubRMSE =0.06 m3/m3) in capturing broad spatial patterns and seasonal dynamics. Beyond enabling accurate large-scale estimates, the explicit formula extracted by KAN clarifies how TB, ST, and VOD influence SM and provides a compact, direct retrieval expression that requires neither KAN retraining nor the iterative solving typical of traditional RTM approaches, thereby improving reproducibility and operational efficiency. This interpretable fusion of physical modeling and data-driven representation offers a novel pathway for advancing quantitative remote sensing retrievals and invites further refinement and operational evaluation.

Preference-based Antibody Expression Ranking: Scaling with Large-scale Weak Supervision

Josh Qixuan Sun, Morteza Babaie, Wenyang Hou, Mark Crowley, and David Young.

In Forty-third International Conference on Machine Learning (ICML). PMLR, Seoul, South Korea. Jul, 2026.

Abs PDF URL

Antibody expression ranking is a critical task in antibody design, yet its modeling is severely hindered by the scarcity of labeled experimental data. To address this, we propose a unified preference-based learning framework that integrates scarce quantitative expression data with large-scale weak positive supervision from immunization data. We adapt Direct Preference Optimization (DPO) to protein language models by introducing a union-masked log-likelihood approximation and IMGT-based alignment, enabling efficient training on variable-length sequences. Evaluating on a diverse internal dataset of 1254 labeled sequences and 4 million unlabeled camelid-derived antibodies, we show that our method consistently outperforms baselines on most metrics. Our results demonstrate that preference learning can effectively learn from weak supervision, providing a scalable solution for antibody expressibility optimization in data-constrained settings.