Professor Javen Qinfeng Shi

Director, Causal AI Group
Areas of interest: Causation, AI, Mind, Metaphysics, and Probabilistic Graphical Models
Professor, School of Computer and Mathematical Sciences (CMS) and Australian Institute for Machine Learning (AIML), University of Adelaide

Email: javen.shi at adelaide.edu.au
Tel: +61-8-8313-0324

Research Mission

We try to understand and causally influence the underlying distributions and processes to help and serve humanity.

Our world is undergoing inevitable and tumultuous changes. Causality, operating beneath the veneer of cause and effect, is essentially the way of change. Our causal AI methods can identify the root causes, discover latent variables, build immunity from spurious correlations, improve generalistion to diverse domains and distribution shifts, model the consequence of interventions, and answer What-If counterfactual questions. More importantly, causal AI holds the key to answer the reverse question: What is the ideal sequence of interventions, given resources or budgets, to optimise future outcomes?

Consulting?

Yes, we do consulting too.

Bio

Professor Javen Qinfeng Shi is the Founding Director of Causal AI Group at the University of Adelaide, and one of the directors at Australian Institute for Machine Learning (AIML). His research interests include causation, AI, mind and metaphysics. Google Scholar ranks him 7th globally in probabilistic graphical models, and 4th in causation. He served as a panellist for the Responsible AI Think Tank from 2022 to 2024 and currently holds the position of an AI Industry Forum panellist from 2024 onward, actively contributing to the cultivation of the national and state AI ecosystem. He has transferred his research to diverse industries including material discovery, agriculture, mining, sport, manufacturing, bushfire, health and education.

Recent awards include: 1) 1st place at Open Catalyst Challenge at NeurIPS AI for Science 2023, using AI to discover energy material; 2) Won AUS/NZ Bushfire Data Quest 2020 using AI to predict fire spread, which led to winning a Citizen Science Grant in 2021, and released a bushfire app NOBURN in 2023 (over 50 media coverages); 3) Finalist of SA Department of Energy and Mining Gawler Challenge 2020 (over 2k participants from 100+ countries) with his team's work being considered as ''The most innovative modelling'' by the judge panel; 4) 2nd place in Explorer Challenge 2019 (over 1k entries from 62 countries); 5) 1st place at SAIC Volkswagen Logistics Innovation Day for Smart Manufacturing 2019.

News

(Aug. 2024) Congrats to Hongrong and Miao on TPAMI 2024 paper A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis, and Recommendation
(Jun. 2024) Congrats to Hongrong and Miao on TPAMI 2024 paper Influence Function Based Second-Order Channel Pruning: Evaluating True Loss Changes For Pruning Is Possible Without Retraining
(Jan. 2024) Congrats to Yuhang etal on ICLR 2024 paper Identifiable Latent Polynomial Causal Models Through the Lens of Change
(Dec. 2023) We won Open Catalyst Challenge at NeurIPS AI for Science 2023. Winners in 2022 and 2021 were Tecent and Microsoft. Congrats to Henry and the team.
(Nov. 2023) Awarded $544,551 from ARC via our discovery project (Ehsan and myself) Learning to Reason in Reinforcement Learning, which aims to equip RL with capabilities such as counterfactual reasoning and outcome anticipation to significantly reduce the number of interactions required, improve generalisation, and provide the agent with the capability to consider the cause-effects.
(Sept. 2023) Congrats to Zhen etal on JMLR 2023 paper Factor Graph Neural Networks
(Aug. 2023) NOBURN app release that assesses bushfire fuel load and reduces the risk (extensive media coverage: Daily Mail, Yahoo News Australia, West Australian, Channel 7 Townsville, Channel 7 Mackay, Triple M (Gold Coast, Darling Downs, Fraser Coast, Wide Bay), Hit FM (Gold Coast, Darling Downs, Fraser Coast, Wide Bay), Canberra Times, Perth Now, Bendigo Advertiser, Newcastle Herald, Yahoo New Zealand, The Examiner, Illawarra Mercury).
(Dec. 2022) Awarded $1.5M ($1M from ARC) for our ARC LIEF (Linkage, Infrastucture, Equipment and Facilities) Grant 2023 titled A Multi-environment Phenotyping Site For Biotech Plants. Congrats to Stu and the multidisciplinary teams.
(Sept. 2022) Post-doc position open on Causality and Re-ID; Salary: AU$$102,952 - $121,779; Duration: 2 years (renewable)
(Sept. 2022) Congrats to Zhen etal on NeurIPS 2022 paper Truncated Matrix Power Iteration for Differentiable DAG Learning
(May 2022) Congrats to Bao, Ehsan and Damith on ICML 2022 paper Bayesian Learning with Information Gain Provably Bounds Risk for a Robust Adversarial Defense
(May 2022) Awarded $2.8M for causality and Re-ID on pest snails control. Congrats to Kym and the multidisciplinary teams.
(March 2022) Congrats to Amin etal on CVPR 2022 paper Active Learning by Feature Mixing
(March 2022) Congrats to Xinyu etal on CVPR 2022 paper Implicit Sample Extension for Unsupervised Person Re-Identification
(March 2022) Congrats to Qingsen and Dong etal on CVPR 2022 paper Learning Bayesian Sparse Networks with Full Experience Replay for Continual Learning
(July 2021) Check the BBC video (4min) that captures our work using AI+IoT+domain expertise to empower Viticulture.
(July 2021) Congrats to Dong etal on accepted ICCV 2021 paper Memory-Augmented Dynamic Neural Relational Inference.
(July 2021) Congrats to Zhenhua etal on accepted ICCV 2021 paper Consistency-Aware Graph Network for Human Interaction Understanding.
(Jun. 2021) Awarded ~$0.5M Citizen Science grant for our research on flighting bushfires in collaboration with USC. See media release at AIML NEWS and AdelaideNow.
(Jan. 2021) Awarded $1M for our research on Causality, Trustworthy AI and Bayesian Deep Net. Congrats to Damith, Ehsan and the team. Hiring.
Awarded $516,000 for our ARC linkage project A Machine Learning driven flow modelling of fragmented rocks in cave mining. Looking for a good postdoc to hire.
(Oct. 2020) Congrats to AIML's $20M direct investment from the federal gov. to boost advanced reasonsing capabilities in AI.
(Oct. 2020) Congrats to Amin and Ehsan on NeurIPS 2020 paper Counterfactual Vision-and-Language Navigation: Unravelling the Unseen.
(Sept. 2020) We entered AUS / NZ 2020 Bushfire Data Quest Challenge. It is great that our technology can help fight the bushfires.
(Aug. 2020) DeepSightX entered ExploreSA: The Gawler Challenge (over 2200 participates worldwide). Check out the pitch video for our solution using AI to predict ternary domains, pseudo geology and mineral distributions.
(July 2020) Our Smarter Regions CRC has gone into Round 1 (we have raised $90M cash from industry and we reuqest $71M cash from the federal gov).
(July 2020) Congrats to Mahsa etal on ECCV 2020 paper Joint Learning of Social Groups, Individuals Action and Sub-group Activities in Videos
(March 2020) Congrats to Ehsan etal on CVPR 2020 paper Gold Seeker: Information Gain from Policy Distributions for Goal-oriented Vision-and-Langauge Reasoning
(March 2020) Congrats to Ehsan etal on CVPR 2020 paper Counterfactual Vision and Language Learning
(Feb. 2020) Congrats to Chongyu etal on TIP paper Model-Free Tracker for Multiple Objects Using Joint Appearance and Motion Inference
(Feb. 2020) Congrats to Yan etal on TKDE paper Fast and Low Memory Cost Matrix Factorization: Algorithm, Analysis and Case Study
(Feb. 2020) Congrats to Qingsen etal on TIP paper Deep HDR Imaging via A Non-local Network
(Feb. 2020) Congrats to Ali etal on IJCAI paper Sparsesense: Human activity recognition from highly sparse sensor data-streams using set-based neural networks
(Feb. 2020) Check out our Smarter Regions Cooperative Research Centre (CRC) Bid. The Smarter Regions CRC will empower regional Australia to gain the maximum benefit from the AI revolution. It will transform existing industries and grow a technology sector in and for regional Australia. Our industry cash commitment (before CRC matching) has surpassed $30M. Together with universities and local governments commitments, total cash after matching is expected to be $80M.
(Sept. 2019) With MIDDOL, we've won Golden Prize (1st place) at SAIC Volkswagen's Logistics Innovation Day with Digital Factory powered by AI! Proud of the team! Great to work with automobile industry.
(July 2019) Congrats to Zhenhua etal on the ICCV 2019 paper New Convex Relaxations for MRF Inference with Unknown Graphs
(June 2019) Proud of the team DeepSightX using AI and geoscience to predict the minerals. Within merely 3 months, from zero to exceptional, we've managed to produce tools that make industry domain experts proud. For those who came to my call, stood by me and fought along side of me, you have my gratitude. For those who know me well, know what that means. The trophies match and reflect your attributes and memorable moments in this journey. The 2nd Prize and $200k for the Explorer Challenge is only a prelude for the real journey to begin. [ media release ]
(May 2019) Congrats to Alireza etal on the IJCAI 2019 paper SparseSense: Human Activity Recognition from Highly Sparse Sensor Data-streams Using Set-based Neural Networks
Check out our CVPR 2019 MOTChallenge Workshop on Tracking and Surveillance
(Feb. 2019) 5 CVPR 2019 papers accepted. Congrats to the team!

Congrats to Ehsan etal on What's to know? Uncertainty as a Guide to Asking Goal-oriented Questions [pdf]
Congrats to Ehsan etal on A Generative Adversarial Density Estimator (oral)
Congrats to Yuhang etal on Variational Bayesian Dropout [pdf]
Congrats to Qingsen etal on Attention-guided Network for Ghost-free High Dynamic Range Imaging
Congrats to Jie etal on RGBD Based Dimensional Decomposition Residual Network for 3D Semantic Scene Completion

(Nov. 2018) We've started a regular Machine Learning Training and Tutorial program.
(Nov. 2018) Congrats to Qingsen etal on the WACV19 paper Multi-scale Dense Network for Deep High Dynamic Range Imaging.
(Nov. 2018) Congrats to Yao etal on the PR paper Efficient dense labelling of human activity sequences from wearables using fully convolutional networks.
(Nov. 2018) Congrats to Zhang etal on the IJCV paper Cluster Sparsity Field: An Internal Hyperspectral Imagery Prior for Reconstruction.
(Nov. 2018) How to turn a two-step adaptive MRF to one-step? Check Suwichaya etal's Signal Processing paper.
(Nov. 2018) Congrats to Yan etal on the TKDE paper Fast and Low Memory Cost Matrix Factorization.
(Oct. 2018) Congrats to Suwichaya etal on the TIP paper An Adaptive Markov Random Field for Structured Compressive Sensing.
(Oct. 2018) Congrats to Dong etal on the TIP paper MPTV: Matching Pursuit Based Total Variation Minimization for Image Deconvolution.
(Oct. 2018) Want a deep neural net to learn and predict orderless labels/outputs? Checkout our Deep Perm-Set paper v.2 here.
(Oct. 2018) Congrats to Ali etal on the MobiQuitous18 paper Deep Auto-Set: A Deep Auto-Encoder-Set Network for Activity Recognition Using Wearables.
(July 2018) Congrats to Jie etal on the ECCV18 paper Seeing Deeply and Bidirectionally: A Deep Learning Approach for Single Image Reflection Removal.
(July 2018) Congrats to Yuhang etal on the ECCV18 paper Deblurring Natural Image Using Super-Gaussian Fields.
(July 2018) Congrats to Ehsan etal on the BMVC18 paper Active Learning from Noisy Tagged Images.
(Nov. 2017) Want a deep neural net to predict SETs (instead of vectors/matrices/tensors)? Checkout our AAAI18 deep set paper here.
(July 2017) Congrats to Dong, Mingkui, Yanning, Anton for our ICCV17 deblur paper.
(July 2017) Check out our tutorial on graphical models on SoCG17 with Chao Chen.
(April 2017) 1 paper accepted to IJCAI17 on higher order inference for graph matching. Congrats to Zhen, Julian, Yong, Wei and Yanning.
(Mar. 2017) Congrats to Dong&Jie etal on the CVPR17 deep deblur paper
(Jan. 2017) 1 Post-doc position now available; Salary: University Level B (AU$94K p.a.); Duration: 2 years (renewable)
(Nov. 2016) 2 AAAI17 papers accepted (1 Oral, 1 Poster)
(July 2016) 1 ECCV16 and 1 CIKM16 papers accepted (1 Poster, 1 Oral)
(Mar. 2016) 4 CVPR16 papers accepted (1 Oral, 3 Posters)