top of page
Jinjin Zhao
AI Researcher & Practitioner
Email:
Research Interest:
Human-Computer Interaction (HCI) with Computer Science, Data Science, Learning Science, and Cognitive Sciences.
Practical AI Applications:
Computer-Supported Collaborative Learning;
Computational Cognitive Modeling
& Knowledge Tracing.
Publications
Publications
Hello! I am Jinjin (double 'gem' :O)
​
I am passionate about understanding & improving the technologies that can aid & brighten our lives. Besides, I am motivated to bring researchers together to contribute collectively to the society we all live in.
​
September 2021
End-to-end Question Generation to Assist Formative Assessment Design for Conceptual Knowledge Learning
Asia Education Technology Symposium (AETS) 2021
April 2021
March 2021
Selected
oral presentation
(Second Author)
Fast, Fair and Private Data Generation
ICLR 2021 Synthetic Data Generation: Quality, Privacy, Bias
January 2021
Best Paper Award
Targeted Feedback Generation for Constructed-Response Questions
​
AAAI 2021 Workshop on AI Education, virtual
August 2020
Interpretable Personalized Knowledge Tracing and Next Learning Activity Recommendation
​
Cold Start Knowledge Tracing with Attentive Neural Turing Machine
​
A novel approach for Knowledge State Representation and Prediction
​
Evaluating Bayesian Knowledge Tracing for estimating Learner Proficiency and guiding Learner Behavior
​
Introducing Alexa for e-learning
​
ACM Learning@Scale 2020, virtual
May 2020
Knowledge Graph semantic enhancement of input data for improving AI
IEEE Internet Computing 2020
July 2019
Collaborative Deep Denoising Autoencoder Framework for Recommendations
ACM SIGIR 2019, Paris
January 2015
An iterative modeling and trust-region optimization method for batch processes
Industrial & Engineering Chemistry Research 2015
August 2014
Iterative optimization for batch processes through online modeling
IFAC World Congress 2014, Cape Town
Talks
Talks
October 2021
AI application in human learning​
​
My Profile at this event
​
Exploring AI across industry, Ai4 2021
30-mins talk
August 2021
NLP research and application in human learning
​
My Profile at this event
​
Applied Natural Language Processing, NLP Summit 2021
30-mins talk
June 2021
A Five-step Thought Process for Self-discovery: From Chemical Engineering to AI Education Science
​
My Profile at this event
​
WomenTech Global Conference 2021
30-mins talk
June 2021
The mystery of human learning and supernatural powers that help understand & improve how humans learn
My Profile at this event
​
WomenTech Global Conference 2021
30-mins talk
April 2021
Conversational AI in human learning in a corporate setting
My Profile at this event
​
RE•WORK CONVERSATIONAL AI & NLP SUMMIT. April 2021
April 2021
The interdisciplinary nature of human learning and technologies that help understand & improve how humans learn
​
My Profile at this event
​
Women in Data Science (WiDS) 2021
40-mins talk
March 2021
ML Technologies for workforce up-skilling
​
1-hour invited podcast @ Certnexus.com
Panel discussion @ Certnexus.com
February 2021
Targeted Feedback Generation for Constructed-Response Questions
30 mins AAAI 2021 AI Ed Best Paper presentation
August 2020
Various papers (talks) at ACM Learning@Scale 2020
Projects
Projects
Responsible AI in learning Ethics, fairness in learning is a must-have for the learning environment we build for all learners. We scrutinize raw data, dig into gradient descent procedure, rectify optimization objectives, monitor post-launch model behavior, and mitigate biases throughout the data and model pipeline.
Discover skill models of learning A skill model of learning, or a cognitive model of learning, is a descriptive account or computational representation of human thinking about a given concept, skill, or domain. It includes both a way of organizing knowledge within a subject area and an account of how humans develop an accurate and complete knowledge of that subject area. However, the cognitive process has been and will be the biggest mystery of humans being the unique species. We urge to design observable experiments and find insights to understand & improve the cognitive process of learning. We progress beyond with innovation and application of its framework and algorithms to find the truth.
Personalized adaptive learning We land in understanding the individual, the learning goal of the individual, the content, the learning context, and providing tailored learning experiences to accelerate their capability development. The challenge lies in how to describe the individual's prior knowledge and their learning goal, representing the learning content, and contextualizing the learning experience to make learning effective, efficient, and addictive(hopefully).
Find insights in inadequate data In fairytale, DATA is the KING that is powerful and reigns over everything. In reality, a fair KING is extremely rare. Data inadequacy, little or imbalanced data, makes it difficult to find insights or improve performance. We employ meta-learning to optimize the gradient descent of gradient descent and deploy generative techniques to reveal what's real in the data.
Build student models of learning A student model of learning is a computational model that describes or simulates how individuals learn through a learning experience. Knowledge tracing is the family of techniques that aim to approximate the knowledge state of the brain, as well as how the knowledge state transits from one another by observing the way individuals learn. Accuracy, reliability, and interpretability of knowledge tracing is the primary research focus that aims to improve the simulation quality of human learning. From statistical regression, hidden markov model, LSTM, to the noble Attention, we advance in algorithms by understanding what's there and then and proposing what can & should be here and now for the better.
Collaborative learning An agent - a system, voice assistant, or chatbot - can not only entertain us, but also help us achieve learning goals. We delve into agent-human interaction practices, design and develop interactive environment for the agent to be a peer, coach, or moderator. We teach agent how to think and behave with experimental design, A/B testing, data analytics, and human-in-the-loop intervention. In return, agent surprises us in ways we haven't imagined and we like.
Social learning We learn by ourselves, with agents, and also from/with other human beings. Learning is happening mysteriously every second as the new synaptic connection constructs in every and each experience. We research into what kind of social learning environment is effective for whom in what learning context at when in developing what skills. The closed experimental loop empowers us to enrich learning theories.
Technologies that guide Instructional Design Data provides a critical perspective in measuring and guiding the instructional design. We leverage the cutting-edge NLP techniques, attention-like Neural Net architecture to assist learning design, in an evidence-based approach. The innovative technologies also enrich instructional design theories and best practices.
Programs
Programs
Cross-disciplinary Mentorship for junior Scientists and Engineers. Running DWDT ('Dancing With Data Troupe' or 'the Derivative of Wonders w.r.t. Time') initiative in my personal time to accelerate their self-development & self-discovery, in areas of ML skill development, research publications, public/impromptu speaking, career development, and every aspect of contextualizing the journey at the company for their growth. As of now, the troupe has 15 members. It is growing joyfully both internally and externally.
Collaborative research and publication. Besides setting up the right environment for passionate individuals to delve, discover, and contribute to science, I collaborate with scientists and engineers internally and externally in areas of knowledge graph, generative modeling, reinforcement adaptive learning, and mathematical proofs for healthcare, education, and things matter to me and to the society. As of now, 10 research papers have been presented in either conference or journal (out of which, 7 as the first author, 3 as the second author). In addition, 3 papers are under blind peer review (2 as the first author, 1 as the second author).
Education & Experiences
Education & Experiences
Master Degree in Mathematical Modeling and Optimization, Zhejiang University
Majored in System Process Control and Optimization in Chemical Engineering (sounds weird, I know). It was a lovely experience during which I found my love in MATHEMATICAL OPTIMIZATION theory - one of the Allfather of emerging ML, AI, and DS techniques.
Software Developer Engineer (SDE), Amazon
Joined Amazon as an SDE after graduation, Someone Does Everything (sadly). It was a wonderful 2.5 years during which I learned to pick up skills quickly as the job required me to do so. The skills range from programming in X-press, python, java, dealing with ambiguity in problem (all kinds of) solving, to thinking big & dreaming wild.
Experiences in Supply Chain Optimization & Retail, Amazon
Great thanks to the complex nature of Supply Chain and Retail business at Amazon. I progressed on how to find the right solution to solve the problem, using the right amount of energy and asking for help from the right people at the right time.
PC member
Program Committee member
TIPCE 2021
AAAI 2021 Workshop on AI Education
HUSO 2021
HUSO 2020
The Seventh International Conference on Human and Social Analytics
The Sixth International Conference on Human and Social Analytics
K-iLKGC21
KGC Workshop on Knowledge-infused Learning
Co-located with The Knowledge Graph Conference 2021
eLmL 2021
The Thirteenth International Conference on Mobile, Hybrid, and On-line Learning
KiML2020
Knowledge-infused Mining and Learning
Co-located with 26th ACM SIGKDD 2020
WebSci2020
12th ACM Web Science Conference 2020
Contact Me
bottom of page