Oleg Zabluda's blog
Tuesday, January 17, 2017
 
"""
"""
Modern video-games like GTAV, present a world where you can test self-driving car AI’s in large complex urban areas replete with realistic roads, weather, pedestrians, cyclists, and vehicles with zero risk or cost. [...] you can quickly run the car through a barrage of safety critical situations, some of which may only occur every several million miles in reality.
[...]
An initial 8-layer neural net with the AlexNet architecture is being made available as well as the dataset it was trained on. Training was done on raw image input from a forward mounted camera regressed against steering, throttle, yaw, and forward-speed control values produced by an in-game AI. This model is able to steer the car to stay in the lane, stop for other cars, and works well in a variety of weather and lighting conditions
[...]
Tips and Tricks

Adding examples of course correction to the training data is crucial. NVIDIA does this by simulating rotation of real-world images, and ALVINN created its own simulation for adding variety to its training. Since we are already in simulation, course correction can be added by stopping recording, steering the car off course, and recording actions and images taken during course correction. This was done at three levels of severity. Levels one and two consisted of driving the car with a previous model for one and two seconds respectively, then recording corrective actions taken by the in-game AI. The most severe level consisted of performing a random action (hard right, hard left, brake, or strong accelerate) and relinquishing control to the in-game AI after 230ms.
[...]
Predicting steering, speed, etc.. ahead-of-time works by adding future targets to the last fully connected layer, but causes more overfitting (e.g. 2x worst test performance with 3 frames, or 775ms, of advance data). So if you predict the future, it's probably a good idea to add more regularization.

Seminal prior work
1988 - ALVINN
2005 - DAVE
2015 - DeepDriving [1]
2016 - NVIDIA DriveNet
"""
http://deepdrive.io/

[1] DeepDriving
https://plus.google.com/+OlegZabluda/posts/8yuuBLmJ4jw
http://deepdrive.io

Labels:


- posted by Oleg Zabluda @ 7:18 PM | permalink | | | technorati | bloglines | feedster
 
"""
"""
The hydrogen anion is a negative ion of hydrogen, H−. The hydrogen anion is an important constituent of the atmosphere of stars, such as the Sun. [...] The ion has two electrons bound by the electromagnetic force to a nucleus containing one proton. [...] Its existence was first proven theoretically by Hans Bethe in 1929. H− is unusual because, in its free form, it has no bound excited states, as was finally proven in 1977 (Hill 1977).
"""
https://en.wikipedia.org/wiki/Hydrogen_anion

Proof that the H- Ion Has Only One Bound State (1976) Robert Nyden Hill,
"""
It is rigorously demonstrated that the nonrelativistic H ion has only one bound state in the fixed (infinite-mass) nucleus approximation with Coulomb interactions only. The H- ion, made up of a, proton and two electrons, has long been known to have one bound state. ' Additional bound states have never been found, but their nonexistence has so far not been proved. The present note provides the proof in the fixed (infinite-mass) nucleus approximation with Coulomb interactions only. The importance of the present result stems from the qualitative difference between the bound-state spectrum of negative ions (of which H- is the simplest example) and the bound-state spectrum of positive ions and neutrals. Negative ions have only a finite number of bound states, for which correlation effects are decisive (H, for example, is believed to have no bound states in Hartree-Pock approximation). Positive ions and neutrals, on the other hand, have an infinite number of bound states.
"""
http://journals.aps.org.sci-hub.bz/prl/abstract/10.1103/PhysRevLett.38.643
http://journals.aps.org.sci-hub.bz/prl/abstract/10.1103/PhysRevLett.38.643

Labels:


- posted by Oleg Zabluda @ 1:01 PM | permalink | | | technorati | bloglines | feedster
 
"""
"""
"The Loco-Motion" is a 1962 pop song written by American songwriters Gerry Goffin and Carole King. [...] The song is notable for appearing in the American Top 5 three times, each time in a different decade: in 1962 by the American pop singer Little Eva,(U.S. No. 1);in 1974 by American band Grand Funk Railroad (U.S. No. 1); and finally by Australian singer Kylie Minogue in 1988 (U.S. No. 3).

The song is a popular and enduring example of the dance-song genre: much of the lyrics are devoted to a description of the dance itself, usually done as a type of line dance. However, the song came before the dance.
"""
https://www.youtube.com/watch?v=j8z34lhKWrQ
https://www.youtube.com/watch?v=j8z34lhKWrQ

Labels:


- posted by Oleg Zabluda @ 10:51 AM | permalink | | | technorati | bloglines | feedster

Powered by Blogger