Andy Harris

  • Student

Submitted

Fighting Fraud with Machine Learning at American Express
Andy Harris
Posted on November 13, 2018 at 2:22 pm
Global financial institutions spend billions of dollars on credit card fraud. American Express is fighting this fraud through investment in machine learning.

Activity Feed

On November 15, 2018, Andy Harris commented on Turning Feelings into Data: Applying Natural Language Processing to Employee Sentiment :

This is a really interesting application of NLP and it’s great to see these types of models applied to People Analytics! I think the concerns you highlighted above make a lot of sense and are common in this space. To the first concern, one approach might be to have clients manually tag previous submissions with sentiments/concerns and use that data for client-specific model training. To the second concern, perhaps Perception could identify submissions that don’t closely match any predetermined concern or sentiment and have them manually reviewed by the client?

I would also propose one additional concern: responses that contain multiple themes or sentiments. I would be interested to learn more about what happens in cases where employees regularly write about one primary theme and another secondary theme. Is Perception able to identify the trend that the second theme is mentioned frequently across posts even if it is never a primary submission topic?

On November 15, 2018, Andy Harris commented on Put Your Money Where Your Mouth Is. Align’s Approach to Additive Manufacturing :

This is a fascinating use of additive manufacturing! I would be interested to know more about what is covered in Align’s 420 patents and what Align will be losing in terms of competitive advantage when they expire. What sort of things are they doing to iterate on their current product (and patent those iterations) so that competitors can’t directly copy them?

I would be wary of getting into a price war. As you mentioned, some people are already 3D printing their own aligners, so it stands to reason that a price war could go all the way to $0 (e.g. open source software that allows users to upload a photo of their teeth and outputs a file that can be 3D printed using any 3D printing service). I would suggest leaning into Align’s partnership with the dental community and focus promotion on the “officialness” of this product over competitors’ products. I would intuitively guess that dentists’ approval of the product would represent value in a consumer’s mind.

On November 15, 2018, Andy Harris commented on Product Innovation Requires Production Innovation: Additive Manufacturing as an Enabler :

Processing equipment production is an interesting use of additive manufacturing (and one I hadn’t previously seen!). I would be curious to know if the company has done much investigation into the cutoff point where it makes more sense to produce this equipment using traditional production methods rather than with 3D printing. In particular, if there is a piece of equipment that is used regularly (and which Lumiena expects to continue using for the foreseeable future), it might make sense to have it produced with sturdier material (though this may be achievable by simply 3D printing with a different material).

On November 15, 2018, Andy Harris commented on Doctors in china leverage machine learning to improve efficiency and effectiveness :

This is a really interesting space for machine learning! I would be curious to know if (in the short-term or medium-term) Nanfang plans to use this technology for human augmentation (e.g. still relying on a doctor to double-check before diagnosis) or replacement (e.g. use the algorithm’s result as the diagnosis). If it’s the latter, I definitely think that their focus should be on improving accuracy. If it’s the former, I would imaging that they could expand into other forms of cancer detection more quickly.

As the technology progresses, I would be interested to know if the technology can evolve to a point where its detection accuracy is better than that of a human.

On November 15, 2018, Andy Harris commented on It’s all open at Wikimedia Foundation :

I had no idea that the Wikimedia Foundation also uses the open innovation process for product development, process improvement, and strategy! I would be interested to see if this is a technique that other companies could also implement.

I really appreciated your thoughts on how to incentivize contributors through the use of a points system like those used on Reddit and StackOverflow. To address the concerns around information validity and/or bias, I wonder if it would be possible to adopt an identity verification system like the ones used on Twitter or Facebook. That way, some content could carry more weight, because it comes from a “verified” source.

On November 15, 2018, Andy Harris commented on Open Innovation at Amazon Alexa: Crowd-sourcing Its Way to Platform Dominance :

This is a great article on how Amazon has been able to leverage open source innovation to quickly develop far more skills for Alexa than other voice assistants. One aspect I found a bit misleading is that—though Alexa devices are the most common smart home devices—Siri is still the most used mobile voice assistant. [1] One way Amazon could continue to grow Alexa is by extending the brand beyond voice interfaces, especially for situations where another communication medium may be preferable to voice (e.g. text input when the user is at work). For example, in addition to its voice interface on iOS and CarPlay devices, the Siri brand is also used to represent caller ID, text prompt autocomplete, and traffic reminders before an event. [2]

[1] Mangis, Carol. “Which Mobile Voice Assistant Is Used the Most?” PC Mag, July 5, 2018. https://www.pcmag.com/news/362260/which-mobile-voice-assistant-is-used-the-most, accessed November 2018.
[2] Apple, “Siri,” https://www.apple.com/siri/, accessed November 2018.