|
Thomas G. Dietterich
@
tdietterich
Corvallis, OR
|
|
Distinguished Professor (Emeritus), Oregon State Univ.; Former President, Assoc. for the Adv. of Artificial Intelligence; Robust AI & Comput. Sustainability,
|
|
|
4.420
Tweetovi
|
198
Pratim
|
21.424
Osobe koje vas prate
|
| Tweetovi |
|
Thomas G. Dietterich
@tdietterich
|
1 h |
|
I use the mass transit directions feature
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
2 h |
|
I often do, when I'm traveling in an unfamiliar place
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
3 h |
|
Looking forward to speaking at this @bigml ML School twitter.com/bigmlcom/statu…
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
11 h |
|
Has anyone verified this story?
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
11 h |
|
Yes of course
|
||
|
|
||
| Thomas G. Dietterich proslijedio/la je tweet | ||
|
Heart of Snarkness
@heartsnark
|
1. velj |
|
A couple of AI cycles ago (i.e., abt 30 years), a practitioner remarked that when AI arrives, it will just be called software/ algorithms. Which is what it is.
|
||
|
|
||
| Thomas G. Dietterich proslijedio/la je tweet | ||
|
AIhub
@aihuborg
|
12 h |
|
Learning to imitate human demonstrations via CycleGAN - ift.tt/36QyWtl pic.twitter.com/fmEtyTV9KS
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
11 h |
|
I agree with your policy point (hype and responsible AI), but the quality of this interaction is much much better than ELIZA.
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
2. velj |
|
Uh, doesn't a solar radiation sensor do the job? The US has thousands of these deployed across the country
|
||
|
|
||
| Thomas G. Dietterich proslijedio/la je tweet | ||
|
Stan
@mcmc_stan
|
2. velj |
|
StanCon 2020 (August 11-14 Oregon State University) registration is now open discourse.mc-stan.org/t/stancon-2020…
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
1. velj |
|
A grant review panel should assess the importance of the research question, the plausibility of the proposed solution, and the likelihood the investigator can execute. Previous work is relevant, but citation counts are lagging indicators
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
1. velj |
|
I’ve never been on a grant review panel that looked at citation counts
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
1. velj |
|
Like all software (not!)
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
31. sij |
|
The authors claim that that arXiv paper is theirs. AAAI needs a way to verify that. Perhaps there could be an arXiv anonymity access key for each paper that authors could give to AAAI that could be used with an arXiv API to reveal (to AAAI) the authors of the arXiv paper
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
31. sij |
|
Any submission to arXiv must come with the authors' names. But those names would be kept confidential for some time period. Now suppose AAAI wants to check a submission for plagiarism. There is high overlap with an anonymous arXiv submission. 1/
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
31. sij |
|
High achievers should be challenged to prove that they merit the extra attention from a moral standpoint. With added resources comes added responsibility. end/
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
31. sij |
|
There is a big risk, though, that those selected for special interventions will be discouraged (e.g., for those given remedial attention) or egotistical (e.g., for those who show they can achieve more). 5/
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
31. sij |
|
I think both motivation and humility come from being shown that, great as your current performance might be, you could be achieving much more. So instead of turning into a smug know-it-all, I was made aware of my deficits and motivated to address them 4/
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
31. sij |
|
Many countries look for talent in sports and allocate coaching to develop those talents. Similarly, I benefited hugely from "tracking" in high school in the US 3/
|
||
|
|
||
|
Thomas G. Dietterich
@tdietterich
|
31. sij |
|
We try to identify students who could benefit from more instructional support. "Meritocracy" isn't the right word for this, but it is crucial for helping disadvantaged students excel 2/
|
||
|
|
||