Phil 2.27.18

7:00 – 5:00 ASRC MKT

  • More BIC
    • A mechanism is a general process. The idea (which I here leave only roughly stated) is of a causal process which determines (wholly or partly) what the agents do in any simple coordination context. It will be seen that all the examples I have mentioned are of this kind; contrast a mechanism that applies, say, only in two-person cases, or only to matching games, or only in business affairs. In particular, team reasoning is this kind of thing. It applies to any simple coordination context whatsoever. It is a mode of reasoning rather than an argument specific to a context. (pg 126)
    • In particular, [if U is Paretian] the correct theory of Hi-Lo says that all play A. In short, an intuition in favour of C’ supports A-playing in Hi-Lo if we believe that all players are rational and there is one rationality. (pg 130)
      • Another form of dimension reduction – “We are all the same”
  • Machine Theory of Mind
    • We design a Theory of Mind neural network – a ToMnet – which uses meta-learning to build models of the agents it encounters, from observations of their behaviour alone. Through this process, it acquires a strong prior model for agents’ behaviour, as well as the ability to bootstrap to richer predictions about agents’ characteristics and mental states using only a small number of behavioural observations. We apply the ToMnet to agents behaving in simple gridworld environments, showing that it learns to model random, algorithmic, and deep reinforcement learning agents from varied populations, and that it passes classic ToM tasks such as the “SallyAnne” test of recognising that others can hold false beliefs about the world
  • Classifier Technology and the Illusion of Progress (David Hand, 2006)
    • A great many tools have been developed for supervised classification, ranging from early methods such as linear discriminant analysis through to modern developments such as neural networks and support vector machines. A large number of comparative studies have been conducted in attempts to establish the relative superiority of these methods. This paper argues that these comparisons often fail to take into account important aspects of real problems, so that the apparent superiority of more sophisticated methods may be something of an illusion. In particular, simple methods typically yield performance almost as good as more sophisticated methods, to the extent that the difference in performance may be swamped by other sources of uncertainty that generally are not considered in the classical supervised classification paradigm.
  • Sensitivity and Generalization in Neural Networks: an Empirical Study
    • Neural nets generalize better when they’re larger and less sensitive to their inputs, are less sensitive near training data than away from it, and other results from massive experiments. (From @Jascha)
  • Graph-131941
    • The graph represents a network of 6,716 Twitter users whose recent tweets contained “#NIPS2017”, or who were replied to or mentioned in those tweets, taken from a data set limited to a maximum of 18,000 tweets. The network was obtained from Twitter on Friday, 08 December 2017 at 15:30 UTC.
  • Back to Basics: Benchmarking Canonical Evolution Strategies for Playing Atari
    • Evolution Strategies (ES) have recently been demonstrated to be a viable alternative to reinforcement learning (RL) algorithms on a set of challenging deep RL problems, including Atari games and MuJoCo humanoid locomotion benchmarks. While the ES algorithms in that work belonged to the specialized class of natural evolution strategies (which resemble approximate gradient RL algorithms, such as REINFORCE), we demonstrate that even a very basic canonical ES algorithm can achieve the same or even better performance. This success of a basic ES algorithm suggests that the state-of-the-art can be advanced further by integrating the many advances made in the field of ES in the last decades. 
      We also demonstrate qualitatively that ES algorithms have very different performance characteristics than traditional RL algorithms: on some games, they learn to exploit the environment and perform much better while on others they can get stuck in suboptimal local minima. Combining their strengths with those of traditional RL algorithms is therefore likely to lead to new advances in the state of the art.
  • Copied over SheetToMap to the Applications file on TOSHIBA
  • Created a Data folder, which has all the input and output files for the various applications
  • Need to add a curDir variable to LMN
  •  Presentation:
    • I need to put together a 2×2 payoff matrix that covers nomad/flock/stampede – done
    • Some more heat map views, showing nomad, flocking – done
    • De-uglify JuryRoom
    • Timeline of references – done
    • Collapse a few pages 22.5 minutes for presentation and questions – done
  • Start on white paper

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s