Share

In our first ever  “informed, but informal” conversation video, Stefan, Nate, and Tony discuss the latest trends in machine learning from the recent DATAx San Francisco conference, get real about how organizations can (but aren’t) using data to drive change and throw a little shade on Google for teaching us bad search habits on how to search poorly. 

You can go to the YouTube video page to jump around the (fairly long, but interesting!) conversation using the extensive timestamps in the video description. If you make it over there and enjoy what you see, be sure to throw us a like and subscribe, and let us know what you think in the comments.

I’ll also put some highlights below the embed in this post.

Highlights from “Let’s Discuss: Machine Learning Applications”

Topics

  • Machine learning strategy and goals alignment (3:00)
  • Improving recommendation engines with machine learning (4:51)
  • Chatbots and machine learning (11:54)
  • Organizational workflow for determining user intent (13:53)
  • Using machine learning to create content variations based on user intent (15:55)
  • Determining proxy metrics to assess mission impact of content (16:17)
  • How non-profits can use machine learning on public data to determine audience needs (17:18)
  • Using AI for coaching, staff improvement and behavior change within an organization (20:57)
  • Using machine learning to inform editorial decisions for website content at scale (21:57)
  • The importance of modular systems (23:57)
  • Democratizing data (28:20)
  • Using data to support, but not dictate organizational strategy (30:52)
  • Where do I start with machine learning? (34:29)
  • The importance of creating structured content and a taxonomy to power machine learning (37:53)
  • Why should I care about machine learning and AI? (39:31)
  • How should I begin investing in machine learning? Small, incremental, with clear objectives and an expectation of culture change (40:50)

Quotes

  • “No matter how deep (you get) into the details of Machine Learning … it all comes back to ‘you have to understand your problem and you have to understand how you’re going to use the results of machine learning” (3:01)
  • “Machine learning isn’t a replacement for the work you need to do, it’s an enhancement of that work” (4:30)
  • “Machine learning algorithms are very good at finding answers to questions, but they are not particularly good at figuring out what question you should ask” (7:23)
  • “Google has taught everybody a bad way of searching, which is just that there’s this bar and you just type whatever into it and it does its best” (12:10)
  • “The analytics are a gatekeeper on top of the algorithm” (17:01)
  • “Machine learning can change what people within organizations spend their time on … and help people make decisions with more confidence” (21:50)
  • “Nothing can future-proof you better than modularizing your setup” (25:07)
  • “Page views and unique visitors … what’s that telling anybody? It doesn’t do anything for the strategy of the organization. It (just) makes everybody feel good” (33:08)
  • “Just solving [the taxonomy problem], which is well within the wherewithal of every organization is a huge step towards being machine learning or AI-ready“ (38:50)
  • “The good news is that everything you do to prepare yourself for machine learning is just good for your organization regardless” (39:10)

Case Studies

  • Hinge dating app uses machine learning to recommend matches (5:27)
  • Uber uses machine learning to help drivers find the right response when contacted (6:32) 
  • Bloomberg uses machine learning to determine tags, personalize and recommend content, and avoid overfitting (7:41)
  • Do Not Pay uses chatbots powered by machine learning to help users resolve parking tickets (10:18)

Resources and links from “Let’s Discuss …  Machine Learning Applications”