Hello! I am an AI researcher and PhD student working on deep learning model explainability at École Polytechnique Fédérale de Lausanne (EPFL). I'm coadvised by Prof. Tanja Käser at the ML4Ed Lab and Prof. Martin Jaggi at the MLO Lab.
Before moving to Switzerland, I worked for two years at Microsoft AI as a lead engineer for the Open Neural Network eXchange project.
My claim to fame (haha) is that I graduated at 20 as the youngest M.S. in Computer Science recipient in UC Berkeley's history. Since then, I've served as a machine learning lecturer for the Berkeley Division of Data Sciences and the University of Washington CSE Department.
I love people, data, and working on exciting problems at the intersection of the two:
Thank you for taking time out of your day to find out what I do with mine!
We present MultiModN, a multimodal, modular network that fuses latent representations in a sequence of any number, combination, or type of modality while providing granular real-time predictive feedback on any number or combination of predictive tasks. MultiModN's composable pipeline is interpretable-by-design, as well as innately multi-task and robust to the fundamental issue of biased missingness. We perform four experiments on several benchmark MM datasets across 10 real-world tasks (predicting medical diagnoses, academic performance, and weather), and show that MultiModN's sequential MM fusion does not compromise performance compared with a baseline of parallel fusion.[Paper + Video] [Pre-Print] [Code] [Poster]
We investigate how bias transfers through an AI writing support pipeline through a large scale user study with 231 students writing business case peer reviews in German. Students are divided into five groups with different levels of writing support (traditional ML suggestions, control group with no assistance, finetuned versions of GPT2, GPT 3, and GPT3.5). Using GenBit, WEAT, and SEAT, we evaluate the gender bias at various stages of the pipeline: in model embeddings, in suggestions generated by the models, and in reviews written by students. Our results demonstrate that there is no significant difference in gender bias between the resulting peer reviews of groups with and without LLM suggestions. Our research is therefore optimistic about the use of AI writing support in the classroom, showcasing a context where bias in LLMs does not transfer to students' responses.[Paper] [Pre-Print] [Code]
Current approaches in human-centric XAI (e.g. predictive tasks in healthcare, education, or personalized ads) tend to rely on a single explainer. This is a concerning trend given systematic disagreement in explainability methods applied to the same points and underlying black-box models. We propose to shift from post-hoc explainability to designing interpretable neural network architectures; moving away from approximation techniques in human-centric and high impact applications. We identify five needs of human-centric XAI (real-time, accurate, actionable, human-interpretable, and consistent) and propose two schemes for interpretable-by-design neural network workflows (adaptive routing for interpretable conditional computation and diagnostic benchmarks for iterative model learning). We postulate that the future of human-centric XAI is neither in explaining black-boxes nor in reverting to traditional, interpretable models, but in neural networks that are intrinsically interpretable.[Pre-Print]
We use human experts to validate explainable AI approaches in the context of student success prediction. Our pairwise analyses cover five course pairs (nine datasets from Coursera, EdX, and Courseware) that differ in one educationally relevant aspect and popular instance-based explainers. We quantitatively compare the distances between the explanations across courses and methods, then validate the explanations of LIME, SHAP, and a counterfactual-based confounder with 26 semi-structured interviews of university-level educators regarding which features they believe contribute most to student success, which explanations they trust most, and how they could transform these insights into actionable course design decisions. Our results show that quantitatively, explainers significantly disagree with each other about what is important, and qualitatively, experts themselves do not agree on which explanations are most trustworthy.[Paper] [Pre-Print] [Code]
We present RIPPLE, utilizing irregular multivariate time series modeling with graph neural networks to achieve comparable or better accuracy with raw time series clickstreams in comparison to hand-crafted features. Furthermore, we extend concept activation vectors for interpretability in raw time series models. Our experimental analysis on 23 MOOCs with millions of combined interactions over six behavioral dimensions show that models designed with our approach can (i) beat state-of-the-art time series baselines with no feature extraction and (ii) provide interpretable insights for personalized interventions.[Paper] [Pre-Print] [Slides] [Code]
We compare five explainers for black-box neural nets (LIME, PermutationSHAP, KernelSHAP, DiCE, CEM) on the downstream task of student performance prediction for five massive open online courses. Our experiments demonstrate that the families of explainers do not agree with each other on feature importance for the same Bidirectional LSTM models with the same representative set of students. We use Principal Component Analysis, Jensen-Shannon distance, and Spearman's rank-order correlation to quantitatively cross-examine explanations across methods and courses. Our results come to the concerning conclusion that the choice of explainer contains systematic bias and is in fact paramount to the interpretation of the predictive results, even more so than the data the model is trained on.[Paper] [Pre-Print] [Slides] [Code]
We tackle the problem of transferability across MOOCs from different domains and topics, focusing on models for early success prediction. In this paper, we present and analyze three novel strategies to creating generalizable models: 1) pre-training a model on a large set of diverse courses, 2) leveraging the pre-trained model by including meta features about courses to orient downstream tasks, and 3) fine-tuning the meta transfer learning model on previous course iterations. Our experiments on 26 MOOCs with over 145,000 combined enrollments and millions of interactions show that models combining interaction clickstreams and meta information have comparable or better performance than models which have access to previous iterations of the course. With these models, we enable educators to warm-start their predictions for new and ongoing courses.[Paper] [Pre-Print] [Slides] [Code]
While transformer-based language models are undeniably useful, it is a challenge to quantify their performance beyond traditional accuracy metrics. In this paper, we compare BERT-based language models (DistilBERT, BERT, RoBERTa) through snapshots of acquired knowledge at sequential stages of the training process. We contribute a quantitative framework to compare language models through knowledge graph extraction and showcase a part-of-speech analysis to identify the linguistic strengths of each model variant. Using these metrics, machine learning practitioners can compare models, diagnose their models' behavioral strengths and weaknesses, and identify new targeted datasets to improve model performance.
Published at eXplainable AI for Debugging and Diagnosis Workshop at NeurIPS 2021.[Paper] [Poster] [Code]
Open Neural Network Exchange (ONNX) is an open standard for machine learning interoperability. Founded by Microsoft and Facebook, and now supported by over 30 other companies, ONNX defines a common set of operators - the building blocks of machine learning and deep learning models - and a common file format to enable AI developers to use models with a variety of frameworks, tools, runtimes, and compilers. Gave several research talks about model operationalization and acceleration with ONNX and ONNX Runtime at Microsoft MLADS (ML, AI, Data Science Conference) and UW eScience Institute.
We present a simple yet effective machine learning model to predict tags for datasets from the United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA) with the labels and attributes of the Humanitarian Exchange Language (HXL) Standard for data interoperability. This paper details the methodology used to predict the corresponding tags and attributes for a given dataset with an accuracy of 94% for HXL header tags and an accuracy of 92% for descriptive attributes. Compared to previous work, our workflow provides a 14% accuracy increase and is a novel case study of using ML to enhance humanitarian data.
A detailed research report on autograding, analytics, and scaling JupyterHub infrastructure highlighted in use for thousands of students taking Data 8 at UC Berkeley. Thesis presented as a graduate student affiliated with RISELab, after helping develop UC Berkeley data science's software infrastructure stack including JupyterHub, autograding with OkPy, Gradescope, and authentication for 1000s of students. Collaborated with Yuvi Panda, Ryan Lovett, Chris Holdgraf, and Gunjan Baid on a talk detailing the infrastructure stack at JupyterCon 2017.
Knowledge Tracing is a body of learning science literature that seeks to model student knowledge acquisition through their interaction with coursework. This paper uses a recurrent neural network (LSTM) and free-form code attempts to model student knowledge in large scale computer science classes.
Working on a framework for deep learning / ML framework interoperability (ONNX) alongside an ecosystem of converters, containers, and inference engines.
Lead of the inter-company ONNX Special Interest Group (SIG) for Model Zoo and Tutorials with Microsoft, Intel, Facebook, IBM, nVidia, RedHat, and other academic and industry collaborators.
Presented and represented Microsoft AI at several conferences: WIDS 2020, Microsoft //Build 2019, KDD 2019, Microsoft Research Faculty Summit 2019, UC Berkeley AI for Social Impact Conference 2018, Women in Cloud Summit 2018, RISECamp 2018
Worked on projects in AI + Systems with an application area of data science education. Project areas include JupyterHub architecture, custom deployments, OkPy autograding integration, Jupyter noteboook extensions, and D3.js / PlotLy visualizations for data science explorations of funding and enrollment data.
Worked on the CSi2 project as a Machine Learning Research Scientist intern on the Hybrid Cloud team. The CSi2 algorithm is an ensemble machine learning algorithm to detect inactivity of VMs as well as suggest a course of action (i.e. termination, snapshot). It is projected to save IBM Research at least $3.2 million dollars with 95.12% recall and 88% F1 score (>> industry standard) and is being implemented into the Watson Services Platform. Collaborated with Neeraj Asthana, Sai Zheng, Ivan D'ell Era, Aman Chanana. Presented an exit talk and filed 2 patents.
Interned at LinkedIn headquarters with the Growth Division's Search Engine Optimization (SEO) Team the summer before entering UC Berkeley. Worked on fullstack testing infrastructure for the public profile pages, as well as a Hadoop project; outside of assigned work, helped plan LinkedIn’s DevelopHER Hackathon and worked on several Market Research/User Experience Design initiatives.
Spent a summer learning computer science fundamentals and shadowing engineers through the CAPE high school internship program at Google Headquarters in Mountain View, CA. Chosen as a Google Ambassador for Computer Science following the experience. Worked with Google, Salesforce, and AT&T to introduce coding to over 15,000 girls across California with the Made w/ Code Initiative.
AIED Program Committee 2023
AIED 2021*, 2022* (Subreviewer for Tanja Käser)
EMNLP BlackBoxNLP 2021, 2022, 2023
Journal of Educational Data Mining (JEDM) 2022
LAK 2022*, 2023* (Subreviewer for Tanja Käser)
Editor for Springer Series on Big Data Management (Educational Data Science)
Fairness Working Group @ EDM 2022
WiML Workshop Team @ NeurIPS 2021
Lead of the 2020 ONNX SIG for Models and Tutorials