Hello, world! I’m Andy. I work at Microsoft. I’m a natural language processing researcher, entrepreneur, VC, and activist. I like to write, but I’d rather build things. I’m fascinated by the ways that technology (when carefully applied) can bring about real impact. I hate injustice.
Things I think I think
I think bringing early-stage technologies to market is hard, but worthwhile. I’m a Principal Group Program Manager at Microsoft working on bringing new natural language processing technologies to enterprises of all sizes. I’m not just a product guy: I’m also part of the exciting NLP research and development work happening here at Microsoft. I previously was the Chief Product Officer for the Intel® Saffron™ Cognitive Solutions Group in Hillsboro, Oregon.
I think success should be measured in terms of global impact. I was a Senior Director for Innovation at Vulcan Inc., where I led teams that tackled some of the world's most important problems, from climate change and clean energy to biodiversity and global health. I co-founded the Vulcan Proving Ground (an early-stage technology incubator) and worked closely with Paul Allen as his Executive Technical Advisor.
I think that small teams can accomplish amazing things. I've founded (and run) three consumer-facing startups: A.R.O. Inc., Swingly, and Extractiv. I also served as CEO of Language Computer Corporation, an amazing natural language processing company that built some of the best systems in question-answering, information extraction, and textual inference in the early 2000s.
I think that hard problems are fun to solve. I've led research in natural language processing, machine learning, artificial intelligence, computer vision, and ubiquitous computing. I even publish (sometimes). My work can be found in the proceedings of AAAI, ACL, SIGIR, and NIPS.
Random headshots from over the years:
An ever-growing list of things I’d like to learn more about:
Neuromorphology, neural structure and function
Neurobiology of decision making, especially reasoning under uncertainty
Artificial Intelligence and Machine Learning
Generative models for language, image, and video
Automated training models based on synthetic sources of data
Computational perception, including computer vision, digital olfaction
Explaining predictions of machine learning-based models
Modeling/anticipating sources of bias
Semi-supervised learning, active learning, machine teaching
Acquisition of common sense knowledge (and other kinds of knowledge)
Natural Language Processing and Linguistics
Recognition of false and/or contradictory knowledge (“fake news”)
Multi-turn dialog agents
Computational approaches to inferring discourse structure
Machine reading and textual inference
Automatic question-answering (including answering complex questions!)
Robotics and autonomy
Required infrastructure for autonomous vehicles in developing world
Quantification of ethical risk / ethical reasoning (“Creepy Trolley”)
Robust strategies for locomotion
Sensor fusion and SLAM
The evolution of cities, and urban planning
Traffic design and transportation planning
Resource planning, especially under unreliable/scarce supply
The evolution of governments and representational systems
Tracking changes in government and policy
Organizational Psychology, especially operations theory, business process redesign
Economic and behavioral incentives, game theory
Innovation at scale
Incentives for sponsoring cutting-edge R&D, opportunity zones
Alternatives to startup incubators, etc.
Policy vs. global pandemics
Gene discovery and editing
Automatic discovery of synthesis pathways
Conservation and biodiversity
Large-scale remote sensing for resource preservation
Ocean health; countering ocean acidification and pollution
Economic incentives (or dis-incentives) for conservation