- Ph.D. University of Illinois at Urbana-Champaign, Theoretical Physics
- B.S. Beijing Normal University, Physics
Amazon Inc.
- Special Projects
Clarifai Inc.
- Computer Vision
- Detection and tracking
- Image embedding
- Fine-grained classification
- Semantic segmentation
- Natural Languange Processing
- Token classification
- Named Entity Recognition (NER)
- DevOps
- Docker
- K8s and Kubeflow
- CI/CD
Insight Data Science
- Satellite image processing
- Partial convolution inpainting model
- Repo and project page
University of Illinois at Urbana-Champaign
- Heat Kernel
- Exact solution of heat equation in a high-dimensional sphere
- Applied to SVM document classification
- Publication and additional proof
- Data Transformation
- Effective Dissimilarity Transformation
- Applied to cell line clustering
- Publication
- Spectral Clustering
- Reinterpreted spectral clustering through the lens of quantum tunneling
- Publication
-
- Customizable and light-weight neural network training toolbox
- Built on PyTorch
- Pronounced "Torch Lighter"
-
- Randomly scattered MNIST digits for quick detector experimentation
- You can also use FashionMNIST or alike
-
- Repo-free experiment containerization
- Build, test, and push
Docker
images all inPython
- Reuse experiement images through
Python
class inheritance
-
Kubeflow
is very much likeTensorflow
(of course they share the last name)Mario
is theKera
orPyTorch
ofKubeflow
- Users can intuitively build pipelines without lengthy and confusing API bureaucracy
-
- Manage and visualize hierarchical concepts in style
- Serialize trees in compact yaml files
- Organize concepts by level and represent concepts by atomic concepts
-
- Perform calculations in various hypothetical scenarios with ease
- Symbolic manipulations based on
Sympy
- Great tool for case study or back-of-envelope calculations
- Almost pure
Python
, highly portable, even on mobile devices, calculation on the go
-
Machine Learning Blog
- Machine Learning Demystified
- My scattered thoughts on ML
Theoretical Physics | Mathematics | Probability | Statistical Learning
PyTorch | Distributed Training | TorchScript
Fine-Grained Classification | Detection | Tracking | ReID | Segmentation
Token Classification | Named Entity Recognition | Transformers
Nvidia Triton | Apple CoreML | Google Protobuf
Docker | Kubeflow | Kubernetes | CI | CD