Deep neural networks (DNNs) have shown to perform very well on large scale object recognition problems and lead to widespread use for real-world applications, including situations where DNN are implemented as "black boxes". A promising approach to secure their use is to accept decisions that are likely to be correct while discarding the others. In this work, we propose DOCTOR, a simple method that aims to identify whether the prediction of a DNN classifier should (or should not) be trusted so that, consequently, it would be possible to accept it or to reject it. Two scenarios are investigated: Totally Black Box (TBB) where only the soft-predictions are available and Partially Black Box (PBB) where gradient-propagation to perform input pre-processing is allowed. Empirically, we show that DOCTOR outperforms all state-of-the-art methods on various well-known images and sentiment analysis datasets. In particular, we observe a reduction of up to 4% of the false rejection rate (FRR) in the PBB scenario. DOCTOR can be applied to any pre-trained model, it does not require prior information about the underlying dataset and is as simple as the simplest available methods in the literature.

Strengths: This paper is quite amazing. Just adding a few linear layers causes otherwise standard, deterministic autoencoders to learn interesting generative factors of a similar or possibly greater quality to VAEs. It's rare to see a simple idea that works very well, with many possible extensions. I think this result will be of wide interest to the community. The theoretical observation that gradient descent dynamics in deep linear networks finds low rank solutions is well established, but has not been put to practical use.

Deep neural networks (DNNs) have shown to perform very well on large scale object recognition problems and lead to widespread use for real-world applications, including situations where DNN are implemented as "black boxes". A promising approach to secure their use is to accept decisions that are likely to be correct while discarding the others. In this work, we propose DOCTOR, a simple method that aims to identify whether the prediction of a DNN classifier should (or should not) be trusted so that, consequently, it would be possible to accept it or to reject it. Two scenarios are investigated: Totally Black Box (TBB) where only the soft-predictions are available and Partially Black Box (PBB) where gradient-propagation to perform input pre-processing is allowed. Empirically, we show that DOCTOR outperforms all state-of-the-art methods on various well-known images and sentiment analysis datasets.

This function will take the sentence scores we generated above as well as a value for the top k highest scoring sentences to sue for summarization. It will return a string summary of the concatenated top sentences, as well as the sentence scores of the sentences used in the summarization. Let's use the function to generate the summary. And let's check out the summary sentence scores for good measure. The summary seems reasonable at a quick pass, given the text of the article. Try out this simple method on some other text for further evidence.

Keras-surgeon provides simple methods for modifying trained Keras models. Keras-surgeon is compatible with any model architecture. Any number of layers can be modified in a single traversal of the network. These kinds of modifications are sometimes known as network surgery which inspired the name of this package. The operations module contains simple methods to perform network surgery on a single layer within a model.

Everybody seems to want to get in on the machine learning hype these days. According to the head of Google's DeepMind team, however, "There's only a few hundred people in the world that can do that really well." No wonder that Gartner has machine learning at the absolute apex of its 2016 Hype Cycle. Nothing else promises to do so much to transform humanity...with such an anemic record to show for itself. This isn't to suggest that artificial intelligence has a bleak future, but rather that we're getting way ahead of ourselves in terms of what it's delivering today.

Categorical variables are known to hide and mask lots of interesting information in a data set. It's crucial to learn the methods of dealing with such variables. If you won't, many a times, you'd miss out on finding the most important variables in a model. It has happened with me. Initially, I used to focus more on numerical variables.

In this paper new hierarchical hybrid fuzzy-crisp methods for decision making and action selection of an agent in soccer simulation 3D environment are presented. First, the skills of an agent are introduced, implemented and classified in two layers, the basicskills and the highlevel skills. In the second layer, a twophase mechanism for decision making is introduced. In phase one, some useful methods are implemented which check the agent's situation for performing required skills. In the next phase, the team str ategy, team for mation, agent's role and the agent's positioning system are introduced. A fuzzy logical approach is employed to recognize the team strategy and further more to tell the player the best position to move. At last, we comprised our implemented algor ithm in the Robocup Soccer Simulation 3D environment and results showed th eefficiency of the introduced methodology.