Improvements to the Neural GPU to reduce training time
Machine Learning, Data Science, Big Data, Psychometrics, Adaptive Learning · Algorithm learning is a core problem in artificial intelligence with… · More significant implications on automation level that can be achieved by machines. Recently deep learning methods are emerging for synthesizing an algorithm from its input-output examples, the most successful being the Neural GPU, capable of learning multiplication.
Item Response Theory allows for ML measuring ability
Machine Learning, Data Science, Big Data, Psychometrics, Adaptive Learning · Item Response Theory (IRT) allows for measuring ability of Machine… · More Learning models as compared to a human population. However, it is difficult to create a large dataset to train the ability of deep neural network models (DNNs). We propose fine-tuning as a new training process, where a model pre-trained on a large dataset is fine-tuned with a small supplemental training set.
science of how people relate to each other to social tech
Developer Advocate · Understanding the science of how 900+ million Facebook users relate with each other in terms of social mitigation, conflict… · More resolution, and compassion as an essential trait in understanding relationships better.
Partnered with researchers from Columbia, Berkeley, and Yale in the field of communicating emotion and social-emotional learning. Led by Arturo Bejar, Director of Engineering at Facebook.
Back End Developer · Hack, a programming language developed for HHVM that interoperates seamlessly with PHP. Hack reconciles the fast development cycle… · More of PHP with the discipline provided by static typing, while adding many features commonly found in other modern programming languages.
UX Designer · Concept #1 d.radio allows you to tune into and listen to d.classes or spaces which are currently broadcasting their audio. You not… · More only listen but are becoming an active participant thru – our secret sauce here
Concept #2 d.mix and match allows teachers, students, professional and anyone to build their own ideal learning experience, based on: - d.school content - external sources: e.g. YouTube, TED Talks, Blogs etc and share it with the world.
The overall problem we tackle is how do we turn passive observers, who just browse the d.website, into active participants, who create, contribute and make the world a better place.
Involved in Typhoon Haiyan relief effort; managed to raise funding to educate out-of-school children. Initiated non-profit agencies with Johns Hopkins for homeless families in Baltimore, outreach meal programs for homeless with UPenn in San Francisco during winter.
Traded futures and Foreign Exchange for Wall Street. Worked at California State Senate and Defense companies. Stanford d.school hack.d Founding Member.
Self-taught guitar/bass/piano player at 13 with music composition, Started coding at 11; and self-taught HTML, CSS, Objective C, and PHP/Facebook Hack programmer. Created FB apps.