Home > Uncategorized > Weekly QuEST Discussion Topics and News, 14 Aug

Weekly QuEST Discussion Topics and News, 14 Aug

14 Aug 2015 QuEST

We have been building a common understanding of the use of CNNs/RNNs for afrl conscious content curation (AC3) incubator:

We started with the review article from Le Cun /Bengio/Hinton  –

from Nature 4 3 6 | N AT U R E | VO L 5 2 1 | 2 8 M AY 2 0 1 5

  • Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.
  • Next we want to hit some related articles to expand the details of the combination of CNNs and RNNs – in particular we want to work towards the models we have up and running “Long-term Recurrent Convolutional Networks for Visual Recognition and Description” by Donahue from UT Austin … et al

– but i want to start with the article that has ‘attention’ as part of its basis:

by Socher et al. Show, Attend and Tell: Neural Image Caption Generation with Visual Attention

  • Inspired by recent work in machine translation and object detection, we introduce an attentionbased model that automatically learns to describe the content of images. We describe how we can train this model in a deterministic manner using standard backpropagation techniques and stochastically by maximizing a variational lower bound. We also show through visualization howthe model is able to automatically learn to fix its gaze on salient objects while generating the corresponding words in the output sequence. We validate the use of attention with state-of-the art performance on three benchmark datasets: Flickr8k, Flickr30k and MS COCO.

Specifically I want to follow the ideas of ‘attention’ in the context of these CNN/RNN combination systems and also focus down on the LSTM – long short term memory networks as a particular instantiation of the RNN piece – I may also have to refer to the article from Socher Improved Semantic Representation from Tree-Structured LSTM networks – as a generalization of the ideas

Where I want to go with this discussion is to hit the models of Donahue et al : Long term recurrent CNN for visual recognition and description (LRCN) that we have functioning processing images/video AND then finally how we intend to make the system more QuEST compliant – specifically if we use the thought vectors as an instantiation of our qualia space (Q-space) how can we enforce our Theory of Consciousness on that representation – so we want to hit our tenets and discuss with respect to that space.

news summary (22)

Advertisements
Categories: Uncategorized
  1. No comments yet.
  1. No trackbacks yet.

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s

%d bloggers like this: