• Qi She


    Intel Labs

    Qi is a Research Scientist at Intel Labs focusing on statistical machine learning, deep learning with applications in computer vision tasks, and robotics. He obtained Ph.D. in machine learning and neural computation from the Department of Electronic Engineering (now Electrical Engineering) at the City University of Hong Kong where he was advised by Prof. Rosa H.M. Chan.


    During this period, He won the 2nd place in the 10th Global Artificial Intelligence Hackathon funded by IBM Watson research. He is also closely collaborated with Prof. Guanrong Chen and Prof. James Kwok, in complex networks and machine learning respectively. He used to be a fully-funded Visiting Student Research Collaborator (VSRC) at Princeton University, advised by Prof. Jonathan Pillow, studying latent structure discovery from high-dimensional neural responses. Before directly pursuing my Ph.D., he completed B.Eng. within 1% (2/230) in Information Engineering from NUPT, and his bachelor final year project ranked #1 out of 230.


    Previously, his research focused on statistical machine learning methods to extract hidden structure from high-dimensional neural data, infer brain connectivity using fully & empirical Bayes, and complex network study on multiple brain regions. Studying how information is encoded, decoded, and processed in our brains is one of his interests. Currently, he is developing a lifelong/continual adaptation agent that can shape a cultivated understanding of the world from the current scene and their previous knowledge via an autonomous lifelong development. The ongoing project is listed in "Lifelong Robotic Vision" project page.


    At the academy, he has more than 20 peer-reviewed publications. He was the organizer of the IROS 2019 Lifelong Robotic Vision Challenge and competition chair of the CVPR 2020 Continual Learning in Computer Vision Workshop, also works as the PC member of ICONIP 2019 and serves as a reviewer for prestigious conferences and journals including NeurIPS, ICML, ICLR, CVPR, AAAI, IJCAI, ICONIP, TSP, EJN etc.


    In the industry, he holds 3 granted/filed US patents. The work is more related to robotics, more broadly autonomous systems using visual learning methods, and developing a continual learning framework/toolkit benefiting quickly prototyping the continual/few-shot/meta-learning applications.

  • Work Experience

    Intel Labs

    Senior Research Scientist, 2018/8-Current

    • Statistical Machine Learning (Graphical model, Generative Model, Nonparametric Bayesian, Adversarial Learning, Few-shot, and Meta-Learning)
    • Computer vision (Multi-modal Action Recognition and Emotional Computing )
    • Lifelong/Continual Robot Learning (Rehearsal based/generative model-based/regularization based)

    Tencent AI Lab

    Research Intern, 2017/12-2018/3

    • Established deep generative models for high-dimensional time series data, aiming at retrieving the intrinsic latent structure of the large-scale dataset.
    • Explored interpretable and intuitive model & tractable and efficient inference method for prediction and visualization of data.
  • Education

    Princeton Neuroscience Institute, Princeton University

    Visiting Student Research Collaborator, 2017/03-2017/12

    • Developed advanced statistical machine learning algorithms for neural data, including stochastic dynamical system, Bayesian inference.
    • Extracted interpretable latent structure of Motor Neuronal Spikings via RNN dynamics and Gaussian Process nonlinear embeddings.

    Department of Electronic Engineering, City University of Hong Kong

    Ph.D. 2014/09-2018/08

    Nanjing University of Posts and Telecommunications

    B.Eng. 2010/09-2014/06

    • GPA: 93/100 (Ranking 2/230)
    • GRE: 320; TOEFL: 110 
    • 1st Prize in National Undergraduate Mathematical Contest in Modeling
    • Excellent Graduation Award (Top 1%)
    • Excellent Final Year Project Award (1st place in the college)
  • Selected Projects

    Contact me if you are interested in any collaboration opportunity! I am always open to exchange the ideas and make potential research happen

    More recent publication updates, please see Qi She's Google Scholar Page

    1st Lifelong Robotic Vision Challenge - IROS 2019 Lifelong Object Recognition Report


    Qi She et al, "IROS 2019 Lifelong Robotic Vision: Object Recognition Challenge"

    published at IEEE Robotics & Automation Magazine, June 2020


    This report summarizes the IROS 2019-Lifelong Robotic Vision Competition (Lifelong Object Recognition Challenge) with methods and results from the top 8 finalists (out of over~150 teams). The competition dataset (L)ifel(O)ng (R)obotic V(IS)ion (OpenLORIS) - Object Recognition (OpenLORIS-object) is designed for driving lifelong/continual learning research and application in robotic vision domain, with everyday objects in home, office, campus, and mall scenarios. The dataset explicitly quantifies the variants of illumination, object occlusion, object size, camera-object distance/angles, and clutter information. Rules are designed to quantify the learning capability of the robotic vision system when faced with the objects appearing in the dynamic environments in the contest. Individual reports, dataset information, rules, and released source codes are also provided.

    CatNet: Class Incremental 3D ConvNets for Lifelong Egocentric Gesture Recognition


    Zhengwei Wang, Qi She* (corresponding author), Tejo Chalasani, Aljosa Smolic, "CatNet: Class Incremental 3D ConvNets for Lifelong Egocentric Gesture Recognition"

    published at CVPR 2020 Continual Learning on Computer Vision Workshop


    Egocentric gestures are the most natural form of communication for humans to interact with wearable devices such as VR/AR helmets and glasses. A major issue in such scenarios for real-world applications is that may easily become necessary to add new gestures to the system e.g., a proper VR system should allow users to add gestures incrementally. We propose a lifelong 3D convolutional framework.

    A Neuro-AI Interface for Evaluating Generative Adversarial Networks


    Zhengwei Wang, Qi She, Alan F Smeaton, Tomas E Ward, Graham Healy, "A Neuro-AI Interface for Evaluating Generative Adversarial Networks"

    published at ICLR 2020 Bridging AI and Cognitive Science Workshop



    Challenges in Task Incremental Learning for Assistive Robotics


    Fan Feng, Rosa HM Chan, Xuesong Shi, Yimin Zhang, Qi She* (corresponding author), "Challenges in Task Incremental Learning for Assistive Robotics"

    published at IEEE access, 2020


    Robotic vision poses new challenges towards applying visual algorithms developed from these datasets because the latter implicitly assume a fixed set of categories and time-invariant distribution of tasks. In practice, assistive robots should be able to operate in dynamic environments with everyday changes.

    OpenLORIS-Object: A Dataset and Benchmark towards Lifelong Object Recognition


    Qi She, Fan Feng, Xinyue Hao, Qihan Yang, Chuanlin Lan, Vincenzo Lomonaco, Xuesong Shi, Zhengwei Wang, Yao Guo, Yimin Zhang, Fei Qiao, Rosa H. M. Chan, "OpenLORIS-Object: A Dataset and Benchmark towards Lifelong Object Recognition"

    accepted into ICRA 2020


    We provide a new lifelong robotic vision dataset (" OpenLORIS-Object") collected via RGB-D cameras mounted on mobile robots. The dataset embeds the challenges faced by a robot in the real-life application and provides new benchmarks for validating lifelong object recognition algorithms.

    Are We Ready for Service Robots? The OpenLORIS-Scene Datasets for Lifelong SLAM


    Xuesong Shi, Dongjiang Li, Pengpeng Zhao, Qinbin Tian, Chunhao Zhu, Jingwei Song, Fei Qiao, Le Song, Zhigang Wang, Yimin Zhang, Baoxing Qin, Fangshi Wang, Rosa HM Chan, Qi She, "Are We Ready for Service Robots? The OpenLORIS-Scene Datasets for Lifelong SLAM"

    accepted into ICRA 2020


    Service robots should be able to operate autonomously in dynamic and daily changing environments over an extended period of time. To accelerate lifelong SLAM research, we release the OpenLORIS-Scene datasets. The data are collected in real-world indoor scenes, multiple times in each place to include scene changes in real life.

    Neural Dynamics Discovery via Gaussian Process Recurrent Neural Networks


    Qi She, Anqi Wu, "Neural Dynamics Discovery via Gaussian Process Recurrent Neural Networks".

    published at UAI 2019 (oral, top 6.8%)



    Dynamics discovery is challenging with high dimensional noisy neural data. Many dimensionality reduction methods have been widely adopted to extract low-dimensional, smooth, and time-evolving latent trajectories. We propose a novel latent dynamic model that is capable of capturing nonlinear, non-Markovian, long short-term time-dependent dynamics called "GP-RNN".

    Neuroscore: A Brain-inspired Evaluation Metric for Generative Adversarial Networks


    Zhengwei Wang, Qi She, Alan F Smeaton, Tomas E Ward, Graham Healy, "Neuroscore: A Brain-inspired Evaluation Metric for Generative Adversarial Networks"

    published at Neurocomputing, 2020


    Evaluating the performance of GANs is still an open and challenging problem. Existing evaluation metrics often require large sample sizes for evaluation and do not directly reflect the human perception of the image quality. We introduce Neuroscore for evaluating GANs, that more directly reflects psychoperceptual image quality through the utilization of brain signals.

    Reduced-Rank Linear Dynamical Systems

    Qi She, Yuan Gao, Kai Xu, Rosa HM Chan​, "Reduced-Rank Linear Dynamical Systems"

    published at AAAI 2018 (accept rate 22%)


    Linear Dynamical Systems are widely used to study the underlying patterns of multivariate time series. However, existing approaches to LDS modeling mostly learn the latent space with a prescribed dimensionality. We propose Reduced-Rank Linear Dynamical Systems (RRLDS), to automatically retrieve the intrinsic dimensionality of the latent space during model learning.

    Evaluating the small-world-ness of a sampled network: Functional connectivity of entorhinal-hippocampal circuitry


    Qi She, Guanrong Chen, Rosa HM Chan, "Evaluating the small-world-ness of a sampled network: Functional connectivity of entorhinal-hippocampal circuitry"

    published at Nature Scientific Reports


    The amount of publicly accessible experimental data has gradually increased in recent years, which makes it possible to reconsider many longstanding questions in neuroscience. An efficient framework is presented for reconstructing functional connectivity using experimental spike-train data in 10 datasets. It is found that the connectivity structure generated by distance-dependent models is responsible for the observed small-world features of the reconstructed networks. Two metrics for quantifying the small-world-ness both suggest that the reconstructed network from the sampled nodes estimates a more prominent small-world-ness feature than that of the original unknown network.

    Neural Dyanamics Modeling Using Kernal Generalized Linear Models


    The main idea is for analyzing the statistical dependencies, causality in neural ensembles, getting insights from data-driven approaches for understanding the underlying brain mechanism.

    Time-varying network modeling using Polya-Gamma augmentation



    In preparation

  • My Hobbies




    起初因为想锻炼身体爱上了这种运动,后来渐渐发现这种运动带给自己的不仅仅是健硕的身体,更多的是精疲力竭后内心的平静,什么都不想的感觉。Swimming is really wonderful !





  • Contact