[HN Gopher] On the Opportunities and Risks of Foundation Models
___________________________________________________________________
On the Opportunities and Risks of Foundation Models
Author : satorii
Score : 18 points
Date : 2021-08-23 17:32 UTC (5 hours ago)
(HTM) web link (arxiv.org)
(TXT) w3m dump (arxiv.org)
| version_five wrote:
| I'd suggest using the original article title. [Edit, it's been
| updated]
|
| Still have to read the article. It's great to see people
| exploring this. From the first "language models are unsupervised
| multitask learners" type papers, i wish there had been more
| emphasis that the various behaviors these models have are
| essentially a side effect of learning some kind of self
| supervision task. A model has been trained to e.g. predict the
| next word given previous words, and we're happy to discover that
| it can be repurposed as a chatbot. And then people find the
| chatbot has some undesirable behaviors, and talk about fairness
| and governance and all that. When the basic point is the model
| was never really trained to do any of that, its just a word
| predictor. Why did you ever think it would be OK to just let it
| run wild on some other task?
|
| All that to say, a big problem in AI/ML is models getting used
| for things they have no business being used for, and them people
| being at best underwhelmed, or harmed or offended by the results.
| The first step should be asking why is this model suitable for
| making the prediction I'm asking it to, and I think closer
| scrutiny on what these "foundation models" actually do is a good
| direction.
| dang wrote:
| (Title changed now. Submitted title was "What is this new AI
| term, foundation models".)
| phreeza wrote:
| To answer the question the original poster apparently had,
| here are the first two sentences of the abstract:
|
| > AI is undergoing a paradigm shift with the rise of models
| (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at
| scale and are adaptable to a wide range of downstream tasks.
| We call these models foundation models to underscore their
| critically central yet incomplete character.
| AlanYx wrote:
| I'm curious about the format/formatting of this paper. There are
| a few visual roadmaps to the various sections and subsections
| throughout the paper, complete with drawings/iconography (clip
| art?). I haven't seen anything like this before in an academic
| paper. Is it something that's becoming popular in certain
| research communities?
| satorii wrote:
| Interesting findings! Not sure about whether it is within
| certain research communities or a broader trend.
|
| But Clip cloud be a good plug-in for nowadays writings/design
| then, something like Clip empowered unsplash.
___________________________________________________________________
(page generated 2021-08-23 23:02 UTC)