There is a rich design space for interacting with enumerative algorithms, and we believe an equally rich space exists for interacting with neural networks. We have a lot of work left ahead of us to build powerful and trusthworthy interfaces for interpretability. But, if we succeed, interpretability promises to be a powerful tool in enabling meaningful human oversight and in building fair, safe, and aligned AI systems (Crossposted on the Google Open Source Blog) In 2015, our early attempts to visualize how neural networks understand images led to psychedelic images. Soon after, we open sourced our code as De...
Music journalist-at-large - blog, reviews & other media.
Citizen of the @indieweb
This made me wonder, when are you an actual ‘citizen’, that is when do you belong to, in or are a part of the Indieweb? Is it when you develop your user page? Is it when you check IRC/Slack Community regularly just because? Or is it when you have a site that has the badge on it? Or is is simple, are you a citizen if you want to be?
This reminded me of Audrey Watters reflections on the Contrafabulist podcast (can’t remember the exact episode) when she wondered when you actually become a New Yorker?
This would also give you access to all of the data you create. We don’t have much control over how our data is used, yet we are also limited in what we can do with it ourselves. Berners-Lee gives the example of fitness-activity data: rather than it being locked up with a company, we should be able to decide whether or not to share this information and with whom. “If you can’t read it, it should be because I’ve decided that you shouldn’t read it – not because our machines won’t talk to each other,” he says. An app on the Solid platform could pull in your own data, plus any that others have shared with you. “[It’s] much more powerful for you as a user, because you can integrate all the data that you have got access to,” he adds.
I am not sure how MIT’s Solid project exactly works, maybe this is what Dave Winer talks about. However, I wonder what a #IndieWeb world would look like? Instead of worrying tagging people on Twitter or Facebook, use Person Tags to notify other users. Syndicate comments across sites. Whatever happens, we need something more than this.
I was listening to a recent episode of RN Future Tense talkÂ about developing a digital construct of ourselves that would exist long after we die. The idea of this virtual self is so that people could ask our opinion long after we die. This is something captured in a few ways in the Black Mirrors series. However, what I was left wondering is whether such virtual selves, based on understanding of the way we think, could sit a standardised tests, such as NAPLAN etc, for us?
It seems that from a number examples shared on online that it is a lot easier to bring about change and transformation in a new school without the supposed baggage of embedded behaviours. This makes me wonder though whether a new school with the ‘best’ teachers as the answer for change is akin to those diets which provide initial success, but are more often than not sustainable in the long term? Those sorts of diets that people follow to loose 10kg for a wedding and then put on 20kg after all the gloss has worn off? Not sure, but I think that we need a more nuanced approach to change. One that celebrates, builds and supports what is in place, rather than looks for solutions on the outside. What about you?