- PyTorch Developer Tools & Libraries
- PyTorch Responsible AI Development Tools & Libraries
- Web and Mobile Applications powered by PyTorch
PyTorch Developer Tools & Libraries
Build a creative, useful and well-implemented tool or library for improving productivity and efficiency of PyTorch researchers and developers. The tool or library must be a machine learning algorithm, model, or an application. It should help with tasks such as debugging, training, model understanding, encryption, deployment or furthering researchers. The submission must be built on PyTorch (submissions using PyTorch-based libraries like torchvision, torchtext, fast.ai, etc.)
PyTorch Responsible AI Development Tools & Libraries
AI developments are improving and integrating with our work and personal lives. It is essential that we, as PyTorch researchers and developers, put our efforts into building tools, libraries, and web/mobile applications that help developing AI models and applications responsibly. These tools, libraries, apps need to support a researcher or developer to factor in fairness, security, and privacy throughout the entire machine learning development process such as data gathering, model training, model validation, inferences, monitoring, and more. Some questions, not limited to, that these tools need to help answers:
- Is my model biasing on a specific group of people based on their race, income, sexuality, nationality, or limited body abilities?
- Do I know my data sets for training and validation are inclusive?
- Is my model development process secure, so it doesn’t put personal information at risk?
- Do I understand the social impact of my model inferences?
- Is there a way to visualize my model to stop any biased inferences?
Web and Mobile Applications powered by PyTorch
Build an application with the web, mobile interface or/and embedded device powered by PyTorch. The submission must be built on PyTorch (submissions using PyTorch-based libraries like torchvision, torchtext, fast.ai, etc.) and have the web, mobile interface or presented as an embedded device so the end users can interact with it.