LESSWRONG
LW

All of Travis Rivera's Comments + Replies

1Alto Clef
You are right. In contrast, without manual optimization, a huge problem is that the neural networks are learning features limited in the training set (like the position of the digit) the do not apply to the testing set. This makes regular DNNs really prone to position shifting. The CNN model, in my understanding, is really more like a work around, which is manually telling the network to separate two kinds of features from the raw pixels--the actual features and their locations--as feature maps, preventing the network being confused by the features shifting between different location. However, this also means CNNs make the assumption that all the features are at the same size of the reception field, thus making it prone to shape/size shifting. If my understanding above is correct, maybe it's possible to design a model, like CNNs, but instead, manually telling the network to separate three features from the raw pixels--features, locations and sizes. This way, maybe it's possible to create an architecture resistant to size/shape/location shifting. As far as I thought, one possible way of doing that is, instead of treating a picture as (raw pixels in DNN)/(raw feature maps in RNN), treating the picture as vectors or even Bézier curves, thus the features extracted, such as the number of closed areas, are no longer depended to any of the fore-mentioned shifts. However, the actual way of doing it is still under my experiment. The above are just naive thought from a beginner in machine learning, and I can't help but wanting to express them. If there's any errors and/or there are already existed matured architectures fit my description above, please let me know so I could improve myself.... Thanks a lot : ).
Frequency Diagram
Travis Rivera*30

I'm pretty sure that the question being answered is "How to find the probability of having a disease if you tested positive for it." I'm observing people interpreting this to mean "What is the accuracy of the test?" which is not the same thing.

Maybe add a bit to distinguish the two questions?

Reply2
What Should Superintelligence Be Programmed To Do
Travis Rivera*10

Assuming there is some way of divining the utility of an individual I think this can be viewed from the lens of a similar problem where you have many agents and you want to combine their utility in some way there is the possibility of just maximizing the average utility vs maximizing the median utility.

The benefits of maximizing the median utility is that the designer is most likely to be part of the median and maybe those that are in the median would have higher utility than those in the average utility maximize world but then this might come at the exp... (read more)

Reply
What Is The Probability That Impeachment Proceedings Will Be Commenced Against President Donald Trump During His First Term
Travis Rivera*130

There have been 3 US presidents where impeachment procedures have taken place against a president. That's ~7% of US presidents. I think that sets a good prior.

Reply4
Correct Credit Tracking Is Very Important If We Want Our Community To Generate New Good Ideas
Travis Rivera*20

I see, so it's more of a "Who said it first" kind of tracking rather than a reputation system like loan credit.

That will incentivize useful ideas. Might also weakly incentivize the generation of ideas that contain no information whatsoever (not unlike the library of babel.) I would guess credit-tracking would be a useful tool for a human to gauge where an idea might originate but would probably take some extra components to make it a correct credit-tracker without opening it up to be gamed for some kind of profit (or status/incentive/ect). So I guess I agree with the claim.

Reply1
Correct Credit Tracking Is Very Important If We Want Our Community To Generate New Good Ideas
Travis Rivera*10

Can you explain what you mean by "correct credit-tracking" and "new good ideas"?

Reply
2Travis Rivera
I see, so it's more of a "Who said it first" kind of tracking rather than a reputation system like loan credit. That will incentivize useful ideas. Might also weakly incentivize the generation of ideas that contain no information whatsoever (not unlike the library of babel.) I would guess credit-tracking would be a useful tool for a human to gauge where an idea might originate but would probably take some extra components to make it a correct credit-tracker without opening it up to be gamed for some kind of profit (or status/incentive/ect). So I guess I agree with the claim.
alexei*120

Anna might have different definitions, but here are mine:

correct credit-tracking

Basically, there is usually one (or a few) strong proponents of an idea when it's first voiced. Those proponents (or someone who picks it up) does a lot of work to explain and popularize that idea. All those people deserve credit, because sometimes other people come and take those ideas and present as basically their own / ideas they just found around.

new good ideas

Ideas that are still correct. For example, that AI safety is important and highly under-prioritized.

Reply2