Xiaoning Qian @ TAMU ECE

I am a professor at Texas A&M University. I am with the Biomedical Imaging, Sensing, and Genomic Signal Processing Group in the Department of Electrical and Computer Engineering (ECE). I am also an adjunct faculty member in the Department of Computer Science and Engineering (CSE). I hold a joint appointment at the Computational Science Initiative (CSI) at the Brookhaven National Laboratory (BNL). Here is my BNL page.

My research interests include machine learning, Bayesian methods, optimal experimental design, signal/image processing, and their applications in life and materials sciences.

Recruiting

We have multiple openings for postdoc and graduate student research assistants (PhD/MS) on developing machine learning and optimization methods in interdisciplinary applications, including bioinformatics and materials science. Applicants with biology/chemistry backgrounds are welcome to send inquiries about bioinformatics positions. Current TAMU students: please feel free to email with CV, transcript, and a brief research statement. [Position description]

To prospective Ph.D. students

I am always recruiting for graduate students. Research Assistantship (RA) positions will be offered to outstanding students who are hard-working and motivated. Interested candidates should email their CVs, transcripts, and brief research statements.
[Note] I apologize but I may not have time to reply to all the inquiry emails. However, I do reply to the students who have applied for admissions to our department if you have all the requested documents in your inquiry emails.

News & Research Updates

We have participated in reviewing AI4Science efforts with researchers across multiple institutes: [Paper][GitHub][Website]
Our efforts on UQ for reduced-order models (ROMs) for solving differential equations have finally been published by IEEE Transactions on Artificial Intelligence!
Congratulations to the collaborating team with Profs. Shuiwang Ji and Xiaofeng Qian for the accepted ICML 2023 papers on AI for computational materials science!
Congratulations to Yucheng and all the coauthors for the accepted AISTATS 2023 paper on uncertainty-aware video hashing!

Congratulations to Randy, Zepeng, and all the coauthors for the accepted ICML 2022 paper on Bayesian continual learning!
We are starting 2022 with one accepted AISTATS 2022 (led by Randy and Shahin), one ICLR 2022 (led by Arman and Ehsan), and two papers (led by Mingzhou and Zepeng) in ICASSP 2022. Congratulations!
Congratulations to Xuan for successful defense of her PhD thesis and the publication of the Cox model with group predictor regularization in Statistics and Computing.
Our new work on efficient Bayesian active learning with GPC for query synthesis got accepted by NeurIPS 2021. Congratulations, Guang and all the collaborators! For the other students with nice works but not accepted this time, pLease do not get frustrated and keep your hard work. Our good works will be recognized sooner or later.
Congratulations to Puhua Niu, who has received the Chevron Scholarship from our department. Also, congratulations for the acceptance of TRIMER to iScience.
Our new work on Bayesian active learning finally got accepted by AISTATS 2021 and ICLR 2021. Congratulations, Guang and all the collaborators! Our group and collaborators also started a new project on automated feature engineering accepted by AAAI 2021, led by Ziyu and Mingzhou. A good way to start 2021 and say bye to 2020.
Our paper on Bayesian relational learning with Ehsan Hajiramezanali, Arman Hasanzadeh, Nick Duffield, and Krishna Narayanan is accepted by NeurIPS 2020.
We have just launched a new website on Objective Uncertainty Quantification.
Congratulations to Ehsan Hajiramezanali, who has received the Chevron Scholarship from our department for the coming academic year.
Guang's recent paper on filtering and experimental design with stochastic differential equations is accepted by IEEE Transactions on Signal Processing. This work is related to our efforts in AEOLUS.
Two papers on 1) Neural Architecture Distribution Search (NADS) with Randy Ardywibowo, Shahin Boluki, Xinyu Gong, and Atlas Wang; and 2) Graph connection dropout learning with Ehsan Hajiramezanali, Arman Hasanzadeh, Shahin Boluki, Nick Duffield, Krishna Narayanan, and Mingyuan Zhou are accepted by ICML 2020. Thanks to all the coauthors for the hard work!
Our recent work on "Hashing using Bernoulli VAE" by Siamak, Shahin and Mingzhang is accepted by UAI 2020. Congratulations!
Two papers, "Semi-Implicit Stochastic Recurrent Neural Networks" by Ehsan and Arman and "ARSM Gradient Estimator for Supervised Learning to Rank" by Siamak and Shahin are accepted by ICASSP 2020. Ehsan and Arman's paper has been selected in the final list of Best Student Paper Award. Congratulations!
Two papers, including "Learnable Bernoulli dropout for Bayesian deep learning" by Shahin, Randy, and Siamak are accepted by AISTATS 2020.
Two papers on dynamic graph analytics --- "Semi-Implicit Graph Variational Auto-Encoders" and "Variational Graph Recurrent Neural Networks" --- with Ehsan Hajiramezanali, Arman Hasanzadeh, Nick Duffield, Krishna Narayanan, and Mingyuan Zhou are accepted by NeurIPS 2019.
The paper "Autonomous efficient experiment design for materials discovery with Bayesian model averaging" with Shahin Boluki, Anjana Talapatra, Raymundo Arroyave, and Edward Dougherty is published in Physical Review Materials (PRM). The story was covered in Science Daily, Phys.ORG, and Analytics Insight.
NSF proposal "Collaborative Research: Combinatorial Collaborative Clustering for Simultaneous Patient Stratification and Biomarker Identification" is funded by the Information Integration and Informatics (III) program.
The paper "Safe feature screening for generalized LASSO," co-authored with Shaogang Ren, Shuai Huang and Jieping Ye, is published in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI).
The paper "BNP-Seq: Bayesian nonparametric differential expression analysis of sequencing count data," co-authored with Siamak Dadaneh and Mingyuan Zhou, is published in Journal of the American Statistical Association (JASA). Using the gamma/beta negative binomial process, we remove sophisticated ad-hoc pre-processing steps commonly required in existing algorithms. Please take a look at our paper and code if you are interested in differential expression analysis of sequencing counts.