On the Embedding of Reliable Deep Learning in the Enterprise
While the barrier of entry in utilizing cutting-edge deep learning algorithms in data-driven companies is severely low, the level to reach operational excellence, stability and auditability when empowering or even replacing core business functions is extremely high. Based on lessons learned from advising some of the world’s largest organizations, this talk will deep dive into some of the common challenges and solutions when embedding AI products in the enterprise. In particular, several real-life use cases will be presented, ranging from deep learning applications for process automation, financial market predictions to distributed intelligent agents advising decision makers. The talk will touch both theoretical as well as technical aspects and features also a snapshot of our own research in the area of natural language processing and adversarial network derivatives. In addition, a few practical guidelines will be given when designing deep learning products requiring corporate on-premise implementation.
Dr. Proissl is a senior quant & data scientist, software engineer, HPC advisor and AI cloud platform lead at Ernst & Young, where he’s designed/built the firm’s private GPU-empowered analytics cloud from backend to web-frontend, developed numerous AI-driven solutions for large organizations and served has an advisor to overcome some of the key challenges when embedding them in business processes. He held managing roles in cross-border audit & advisory engagements and leading international research collaborations with contributions to AI research, Cognitive Systems and Particle Physics. Previously, he played a key role at CERN to search and discover the Higgs Boson as well as at Brookhaven National Lab to build massively parallel adaptive control & monitoring systems based on Manifold Embedding and Reinforcement Learning concepts to dynamically govern particle detectors for studies of the origin of the universe. He holds a Ph.D. in Particle Physics with over 15 years of R&D experience, developed and embedded software systems in corporate, research and governmental institutions and has published numerous peer-reviewed research.