I am a 5th year PhD candidate in the ACO program (Algorithms, Combinatorics and Optimization) at the Tepper School of Business, Carnegie Mellon University. I am fortunate to be able to work with R. Ravi, Andrew Li and Sridhar Tayur.
Prior to joining CMU, I received my BS degree in Mathematics from Tsinghua University in 2014, and MS degree in Applied Mathematics and Statistics from the State University of New York at Stony Brook in 2017, where I was fortunate to work on computational geometry with Joseph Mitchell and Jie Gao.
I am currently on the job market. I will be giving a talk titled Conservative Price Experimentation: Markdown Pricing Under Unknown Demand in the INFORMS meeting at 7:45-9:15 AM (PST) on Oct 26, in CC - Room 201A (in person) or Virtual Room 26 (virtual). You may also find my 20-minute video on Youtube. Detailed INFORMS schedule can be found here.
Below is a 7-minute video about my recent research:
My research interest is prescriptive analytics and revenue management, mainly on sequential decision-making, statistical learning theory, and approximation algorithms for computationally challenging operations problems.
On the practical side, I am fortunate to be able to collaborate closely with industry to bridge the gap between theory and practice, while exploring the aspects that theoreticians tend to overlook. Currently, I am leading a joint project with Glance, India's largest mobile lock-screen content platform, to improve their recommender system by incorporating ideas from bandits theory. I am also working with Bestar&Bush, a North American furniture manufacturer, on modernizing their global supply chain by leveraging their massive data.
Su Jia, Andrew Li, R. Ravi
Our previous work showed the first separation between markdown and unconstrained pricing, under minimal assumptions. Is there still a separation if we add more assumptions, for example assuming the demand model has certain functional form? We introduce a complexity index that measures the complexity of a family, and provide a complete settlement of the problem under this framework by proving tight regret bounds for each regime. (Click here to watch my informs talk video)
Kyra Gan, Su Jia, Andrew Li, Sridhar Tayur. (NeurIPS'21) Winner, INFORMS Pierskalla Award 2021
Liquid biopsy is an emerging approach for early cancer detection, which aims at detecting mutations from the free-floating DNA in the blood. We consider an Active Hypothesis Testing problem with the goal of identifying the cancer type using minimal number of tests. Apart from theoretical guarantees, we also demonstrated the efficacy of our policies on real DNA mutation data.
Su Jia, Jeremy Karp, R. Ravi, Sridhar Tayur. (Forthcoming, Manufacturing and Service Operations Management)
Retailers nowadays may use in-store inventory to fulfil the demands from different channels. Different from offline orders, online orders may be fulfilled periodically, and thus the key decision is the number of online orders to accept. We propose a gradient-based computational framework and demonstrated its effectiveness on Onera's data.
Su Jia, Andrew Li and R.Ravi.Egon Balas Award for best CMU student OR paper
Dynamic pricing with unknown demand has been extensively studied and often formulated as a bandit problem. While well-understood theoretically, bandit-based policies are rarely deployed in the real world, since many of them overlooked practical constraints. For example, the price may oscillate, which is unfavorable in practice. We consider markdown policies, i.e. policies with non-increasing prices, and show a tight regret bound that "separates" markdown and unconstrained pricing.
Su Jia, Fatemeh Navidi, Viswanath Nagarajan and R.Ravi. (NeurIPS'19)
In medical diagnosis, a sequence of medical tests is performed to identify the patient's disease. We consider the problem of finding the lowest-cost decision tree. As opposed to prior work where the outcome of each test is assumed to be known, here we consider the setting where some tests may have unknown outcomes, and provide the first approximation algorithms.