Ideas that worked. Ideas that didn’t, Ideas that published. Ideas that didn’t publish but had impact. I will focus on the road of TVM from a research project to open source package to community to foundation of a company.

When we participate in research with consequences, it’s important to reflect on how those consequences can have far reaching impacts, for better and for worse. This talk investigates important considerations for ethical research practice and reflections on available mechanisms for accountability.

In this talk I will share my perspectives on the entire process of conceptualizing visionary research ideas, conducting and prototyping them and then of creating actual products out of them in an industrial setting. With various real-world examples I will share the missteps encountered and the learnings garnered from them, which have helped to shape my career.

Text classification is one of the fundamental tasks in natural language processing (NLP), and recent deep learning models have made great progress in this area. However, the text features obtained by common NLP models such as Transformers, textCNN, etc., suffer from a high degree of anisotropy which degenerates the expressiveness of the learned features. Maximal Coding Rate Reduction (MCR2) principle is to maximize the coding rate difference between the whole dataset and the sum of each individual class which can lead to more isotropic representation. We try to migrate MCR2 principle from image classification to text classification. The results show that applying MCR2 principle enables models to obtain more uniform and inter-category orthogonal text embeddings, but at the same time reduces the accuracy of text classification.

In this paper, we introduce a method for dealing with imbalance in data based on SMOTE oversampling and GAN generated images. We assess the change of performance across different methods and discuss the results of the experiments. We also take an attempt of explaining why this method did not succeed to open the discussion on GAN-based resampling methods.

As machine learning (ML) systems take a more prominent and central role in contributing to life-impacting decisions, ensuring their trustworthiness and accountability is of utmost importance. Explanations sit at the core of these desirable attributes of a ML system. The emerging field is frequently called “Explainable AI (XAI)” or “Explainable ML.” The goal of explainable ML is to intuitively explain the predictions of a ML system, while adhering to the needs to various stakeholders. Many explanation techniques were developed with contributions from both academia and industry. However, there are several existing challenges that have not garnered enough interest and serve as roadblocks to widespread adoption of explainable ML. In this short paper, we enumerate challenges in explainable ML from an industry perspective. We hope these challenges will serve as promising future research directions, and would contribute to democratizing explainable ML.