Upload your resume. PhdFit's six research agents compare your background with faculty profiles, recent publications, lab focus, and outreach opportunities, then rank professors with evidence you can review.
Resume-aware matchingEvidence from recent researchOutreach drafts included
Advisor Match Brief Resume uploaded
Resume signals5 found
NLPHCIFairnessGraph MLPyTorch
Six-agent analysisRunning
Axel · Candidate AnalystResume parsed
Nova · Professor ResearcherRecent papers
Echo · Match ExplainerEvidence linked
Atlas · Strategy AdvisorFit ranked
Lyra · Outreach WriterDraft queued
Juno · Meeting Prep CompilerNext step ready
SC
Dr. Sarah Chen
Stanford · NLP & interpretability
91
fit
Fairness + NLP projects overlap with her recent work.
Strong methods fit: model evaluation and interpretability.
Clear outreach angle for current trustworthy AI research.
Outreach angle
Ask how her lab is extending interpretability methods into fairness audits for real-world AI systems.
Zhenong Jin
· Adjunct ProfessorVerified
University of Minnesota · Department of Community Development
Zhenong Jin is an Associate Professor at The Institute of Ecology at Peking University. His research focuses on mapping agricultural features using high-resolution satellite imagery, forecasting crop yields, integrating crop models with remote sensing for precision nitrogen management, and studying the impacts of climate change on agroecosystems. His work involves digital and precision agriculture, employing knowledge-guided machine learning and remote sensing technologies to improve understanding and management of agricultural systems.
Nitrate is a water pollutant with significant environmental and health effects. Current field-deployable nitrate sensors are expensive, limiting monitoring and management efforts. We designed a lower-cost optical nitrate sensor suitable for field use. We built a prototype unit for less than $800 and tested the unit in a laboratory setting in both distilled and river water with different levels of added nitrate. Our calibration shows that the device effectively senses nitrate in the 0 to 100 mg/L range we chose for relevance to real-world applications. We developed a method for temperature correction of the optoelectronics and applied a third-order polynomial fit for nitrate response to reach similar performance in clean water to commercial sensors, with an RMSE of 0.67 mg/L and R2 > 0.99 on the fully averaged data set, and a limit of detection of 1.73 mg/L on 20 min averaged data. We identified a need for an improved reading protocol and further calibration work to address challenges with matrix effects in dirty water. We pointed out the key issues for future production of the sensor, including unit-to-unit variability and interference from other compounds. This novel sensor design provides promise as a low-cost but effective nitrate water quality sensor.
This Zenodo record provides the processed data and scripts used to support the analyses and generate the figures for the associated manuscript. Contents Data/: processed (derived) datasets used in the main analyses. Scripts/: scripts for data processing, analysis, and figure generation. How to reproduce Download and extract the archive. Create the computing environment. Run scripts in Scripts/. Expected outputs: figures corresponding to the main and supplementary figures reported in the manuscript.
Remote Sensing of Environment · 2026-02-05 · 4 citations
articleOpen accessSenior authorCorresponding
Accurate large-scale crop yield mapping is crucial for understanding how weather and climate variability affect food security. While temporal deep learning models have achieved notable success in extracting time-series features for yield mapping, they often struggle to model spatial dependencies, such as yield spatial autocorrelations and the influence of time-invariant variables (e.g., soil properties and topography). To address this limitation, we propose KGML-Graph, a knowledge-guided graph machine learning framework that integrates spatial learning with temporal structures to explicitly capture these underutilized spatial dependencies. We incorporate knowledge-guided edge weights, derived from historical yield correlations, into the graph structure to identify patterns among counties with similar yield dynamics and enhance model training. The framework was evaluated using data from 627 counties across the U.S. Corn Belt from 2000 to 2020, benchmarking against Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and Temporal Convolutional Neural Network (TempCNN). Results show that KGML-Graph outperformed benchmarks in cross-year testing, reducing RMSE by at least 10.8%, and improved R 2 by at least 9.3% in temporal extrapolation during 2017–2020. Furthermore, it demonstrated superior spatial transferability by maintaining accuracy in unseen regions and significantly reducing spatial autocorrelation in estimation residuals. Under extreme climatic conditions, the model achieved at least a 14.4% improvement in R 2 on out-of-distribution test data and reduced the mean estimation residual from -0.413 to -0.074 metric tons per hectare, effectively mitigating the systematic yield overestimation observed in baseline models. Our attribution analyses highlighted the contributions of both graph structure and knowledge-guided edge weights to the improved performance by better capturing spatial patterns and representing key static variables, such as soil organic carbon content. These findings underscore the potential of KGML-Graph as a robust framework for unifying spatial and temporal learning to support accurate, large-scale crop yield mapping across diverse climatic and geographic conditions. • Temporal learning models often overlook spatial dependencies in yield mapping. • We propose KGML-Graph to enable explicit spatial learning from geospatial data. • Knowledge-guided edge weights from yield similarity enhance spatial representation. • KGML-Graph outperforms temporal models in spatiotemporal extrapolation tasks. • Interpretability analyses reveal KGML-Graph captures key time-invariant features.
Repository KITopen (Karlsruhe Institute of Technology) · 2026-01-01
articleOpen access
Accurate estimation of surface soil moisture (SM) in terrestrial ecosystems is essential for understanding hydroclimate dynamics. The L-band Soil Moisture Active Passive (SMAP) mission provides 9-km global daily surface SM by using a microwave radiative transfer model (RTM)-based algorithm. However, the accuracy of SMAP SM is limited in regions with dense vegetation cover and complex surface conditions, due to the empirical parameterization and oversimplified radiative transfer processes. To overcome the limitations, we developed a Process-Guided Machine Learning (PGML) framework to integrate RTM theories and deep learning to predict global daily surface 9-km SM from April 2015 to June 2025. Informed by domain knowledge, we developed the PGML model structure using RTM and hydrological theories, designed a Kling-Gupta efficiency-based cost function, pretrained it with RTM simulations, and fine-tuned it with in-situ measurements. The independent validation shows that PGML SM has strong agreement with in-situ measurements (R = 0.868 and unbiased RMSE = 0.054 m 3 /m 3 ). This study highlights the potential of PGML to enhance the accuracy of satellite SM, thereby supporting improved water resources and ecosystem management.
Zenodo (CERN European Organization for Nuclear Research) · 2026-01-27
datasetOpen access1st authorCorresponding
Source data for Figure 2b&c in the study "Knowledge-guided artificial intelligence in global change ecology research" to be published in Global Change Biology. The table contains two sheets, corresponding to the source data of Fig. 2b and 2c. They respectively come from Liu et al. (2024): "Knowledge-guided machine learning can improve carbon cycle quantification in agroecosystems" (https://doi.org/10.1038/s41467-023-43860-5) and Khandelwal et al. (2025): "Physics Guided Machine Learning Methods for Hydrology" (https://doi.org/10.48550/arXiv.2012.02854).
International Journal of Applied Earth Observation and Geoinformation · 2026-02-17
articleOpen access
• Global analysis reveals salinity regulates VPD constraints on mangrove photosynthesis. • Coupling Sentinel-2 red-edge with climate and salinity data quantifies mangrove responses. • Dry-climate and high-salinity mangroves are most vulnerable to future warming and drying. Atmospheric drought stress limits mangrove photosynthetic activity, and this constraint can be further amplified by high salinity, yet their combined global effects remain poorly understood. Here, we integrated multi-source Earth observation and geoinformation datasets, including Sentinel-2 red-edge position (a proxy for canopy photosynthetic activity), vapor pressure deficit from TerraClimate, seawater salinity from Copernicus reanalysis, to investigate how salinity regulates the sensitivity of mangrove photosynthesis to atmospheric drought stress during 2019–2023. Datasets were harmonized and analyzed through reproducible geoinformation workflows at 10 m–0.5° resolutions, enabling large-scale coupling analyses between remote sensing proxies and climate drivers. We found that drought stress constrained mangrove photosynthetic activity worldwide, with stronger limitations in tropical savannahs than in tropical rainforests. Marine mangroves exposed to persistent high salinity were more sensitive than estuarine mangroves influenced by freshwater inflow. These results reveal a global pattern in which salinity amplifies atmospheric water constraints on mangrove photosynthesis. Mangroves in dry climates and high-salinity habitats are therefore most vulnerable to future warming and drying. Our findings confirm that integrating multi-source satellite observations with geoinformation analysis provides an effective, large-scale approach for assessing vegetation vulnerability and identifying conservation priorities in climate-sensitive mangrove ecosystems.
<h3>Abstract</h3> Cash crops can boost local economies but compete with forests for land. We quantify this tension in Benin using variation in cashew cultivation driven by heterogenous local responses to global cashew price volatility. We develop high-resolution cashew maps using a deep learning model trained on field data and pair it with gridded GDP data. Cashew cultivation degrades forests without generating detectable income gains. Muted income effects reflect (i) concentrated benefits obscured in coarse GDP data, and (ii) farmers valuing income-smoothing benefits of cashews over first order gains. Costbenefit calculations show each dollar earned from cashews incurs $16 in ecological costs.
Zenodo (CERN European Organization for Nuclear Research) · 2026-01-28
datasetOpen access
This Zenodo record provides the processed data and scripts used to support the analyses and generate the figures for the associated manuscript. Contents Data/: processed (derived) datasets used in the main analyses. Scripts/: scripts for data processing, analysis, and figure generation. How to reproduce Download and extract the archive. Create the computing environment. Run scripts in Scripts/. Expected outputs: figures corresponding to the main and supplementary figures reported in the manuscript.
