Applied AI in Finance
I will mostly be focusing on using bayesian methods together with vectors for word representation algorithms (eg: glove, word2vec like algos) to analyze data feeds for better decision making in an applied real world setting. This will be showcased by our ANN’s capability to disseminate unstructured topical information to output an analyst report and a near real time process by which the AI takes a decision based on the input to autonomously invest/trade in a particular financial security. Towards the end, I will also be speaking a bit more about the need to create and support open data protocols for personalization by parsing existing user datasets and allowing them to be freely imported across different types of AI agents.
Shaunak Khire, is the cofounder & CEO of Emma/MANSI AI, a company focused on building autonomous vertically integrated machine intelligence platform using bayesian methods & vectors for representation of words. Emma, the AI’s moniker, called the short term bottom on crude oil and the broader commodities complex in January. It began analyzing and writing analyst reports on US equities and bonds in March. In June 2016, Emma/MANSI began investing autonomously in in three financial securities GSK, Tesla and the 10Yr Bond.
Shaunak started out as a mobile security hacker at BT focusing on vulnerabilities in the Bluetooth standards, specifically the A2DP profiles on the Symbian and Java OS platforms. Previously Shaunak was the interim CTO & Board Member of a private equity backed DMP (data management platform), he has also served on the global board of the Mobile Marketing Association. Prior to that he was a fund manager at a small family office managing a diversified energy portfolio. Shaunak was a part of the Clinton Global Initiative tech & poverty alleviation groups and he coconstructed the MaghaCGI30 Index a social impact index that weighs constituents based on qualitative characteristics found typically in unstructured data such as corporate CSR reports. He is a contributor for TechCrunch & Recode and has been on MIT Technology Review’s 35 under 35 list in 2012.