Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Cracking open the black box of automated machine learning (news.mit.edu)
101 points by espeed on June 1, 2019 | hide | past | favorite | 9 comments


I initially read the headline as "cracking down on the..." And I thought to myself, finally, people are writing about the blind use of machine learning and the scarcity of math and CS knowledge of the average user.

I suppose the actual headline is still a step in the right direction. People need to understand why these things work or not .


That was my first thought as well. In reality, this just allows a bit of insight into the black box of AutoML. It won't provide much insight into the ML black boxes that AutoML searches over. What is really needed are effective ML algorithms that are more transparent and predictable to a wider class of practitioners.


Summary: Describes an interactive tool for monitoring and visualizing an automated search for which of several machine learning algorithms might perform best on a particular dataset.


The tool seems really good for learning model. I am excited. But is it available or open sourced?

I couldn't find it with simple google search?



Hey Dawg, I’ve heard you like AutoML, so I developed a tool to manually tune AutoML.


Researchers from MIT and elsewhere have developed an interactive tool that, for the first time, lets users see and control how automated machine-learning systems work.

That's a really bad way to formulate this. If I were a researcher from "elsewhere" I would feel that my MIT coauthors are dismissing my contribution and are taking credit for my work.


The news office writes the press release not the co-authors or research group, but yeah it kinda sucks. They get a proper shout-out in the third paragraph. Should have really just been written that way in the sub-headline and introduction.


this is mits pr page. every college does the same thing to build the brand of the college. I'm sure the co-authors schools omit MIT in their press release




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: