Intel, Deep Learning Analytics and In-Q-Tel Discuss the Deep Learning Revolution at Jan. 31 Titans

On Jan. 31, more than 400 members of the region’s technology community gathered to discuss the promises and pitfalls of deep learning at NVTC’s Titans event featuring Intel’s Chief Data Scientist, Public Sector, Americas Melvin Greer and Deep Learning Analytics’ Managing Partner and Data Scientist Dr. John Kaufhold. Dr. Ravi Pappu, Chief Architect of In-Q-Tel, moderated.

Kicking off the event, Dr. Pappu shared the history of deep learning and described how it is becoming more accessible for businesses of all types. He highlighted enablers of deep learning, such as algorithms, data, hardware and talent, as well as pitfalls such as errors, bias and transparency.

Kaufhold and Greer emphasized how deep learning is enabling businesses across every sector to solve important problems more efficiently and more cost-effectively. Greer highlighted the importance of marrying the right algorithms, data and hardware to the problem a business is trying to solve. Kaufhold expressed his belief that, if you have the data, AI and deep learning can be applied to almost any problem to make the process cheaper or faster. The panel also highlighted some of the latest advancements being driven by deep learning, including personalized medicine, cyber intelligence, autonomous vehicles and speech recognition.

From left: Intel’s Dr. Ravi Pappu, Deep Learning Analytics’ Dr. John Kaufhold and Intel’s Melvin Greer participated in a panel on deep learning at NVTC’s Jan. 31 Titans event.

The panelists pinpointed some of the impediments that exist today in harnessing the full power of deep learning. Kaufhold expressed his belief that talent and data are the critical gatekeepers for successful applications. Greer agreed that having clean data that is appropriately labeled is crucial and can be the heavy lifting in any deep learning effort. Greer also noted the legal, ethical and societal challenges of deep learning. Both called attention to the widening data science talent gap, the slow adoption of advanced data science curriculum in higher education and how this is posing significant challenges to deep learning advancements and adoption.

During a question and answer session with the audience, panelists addressed concerns about efforts to develop standards in labeling metadata and how to improve trust in deep learning conclusions.