Growing Clojure in AI via the InferenceQL and OpenGen Probabilistic Programming Platforms
How can AI applications best leverage the simplicity, portability, and expressive power of Clojure? And how can the Clojure community draw on the value and significance of AI applications, to grow a stronger, more impactful, and more inclusive community? This talk will show how Clojure/Script is exceptionally well-suited for enterprise AI applications of probabilistic programming, an AI programming model pioneered in part by MIT's Probabilistic Computing Project that has recently been shown to outperform deep learning in terms of data efficiency, robustness, and explainability. It will highlight two Clojure/Script-based probabilistic programming platforms: 1. InferenceQL, a probabilistic programming platform suitable for interactive AI applications and embedded use in enterprise data pipelines. Applications include synthetic data generation and quality assessment, Bayesian AutoML, and analyses of demographic and compensation data for software engineers, to assess diversity, equity, and inclusion challenges. 2. Financial modeling via custom probabilistic programs written in Gen.clj, a Clojure/Script implementation of MIT's Gen probabilististic programming language that is part of the emerging OpenGen platform. I will also briefly discuss opportunities to help grow the impact of Clojure for DEIJ efforts in the software industry more broadly.