KEYNOTE: Do we really need more data?
Once upon a time working on artificial intelligence or machine learning meant constantly yearning for data and struggling to find cool problems to work on for which there was some. Now data’s aplenty but that old longing for data seems to have left a deep scar. And so, most take for granted intelligence is measured in petabytes. True, machine learning works on data. But data isn’t like gold – it isn’t the new oil either. No matter the analogies, harvesting, mininig, piping, maintaining, cleaning and preparing the data requires continual, sophisticated engineering, legal, and business development work, and all you get is data. The value producing part of machine learning is in translating business questions into algorithms, not in the data. So will more data by itself make the machine intelligent? your business more effective? efficient? valuable? Many companies intent on developing an AI strategy end-up falling short of their goal because just getting the data takes over all the strategy. So, is acquiring more data what your organization needs? Is that the path to an AI business? You got the data, then what? much ado about …? I too have experienced the longing for more data, and in every single instance – ranging from telecommunications to pharma— the breakthrough came from finding ways to work with less data—the relevant data— and refining the question we need the machine to learn about from it. What if making AI business-ready for you was altogether different than collecting data? What if it had to do with the (business) questions you ask? That sounds like a conversation worth starting.Introduced by: Kathleen Walch, Managing Partner & Principal Analyst, Cognilytica
Session ID: KEY1315 Presentation Type: Live Keynote Session (Replay Available)
Date / Time: [Day 2] Tue. Sep. 15, 2020 @ 13:00 ET (US)
To view this session, register for the conference.