Goto

Collaborating Authors

P2P or Master-Salve Chatbots Architectures ? -- ConvComp2016

#artificialintelligence

Interesting points; I agree and we already exchanged few tweets about a futuristic vision of a bot-to-bot (sort of p2p) converse (I hypotized a return of something like web services (…failed architecture, I admit) for bots "cooperative" machine to machine services: What you call Meta Chatbot -- A chatbot'router' (and I'd call "master-slave", just joking) appears to me an architecture more faisible now; in a sense, maybe Amazon Alexa is exactly in this way, with Skills plug-in ecosystem: Concerning natural languages conversational software (aka chatbots), and specifically about vertical versus General AI: I agree about the fact many companies claim to supply some sort of mysterious "AI/machine learning" engines (proprietary cloud platforms often), but until now is not to clear (to me) if these software contains real machine learning (or are just baeysian clasifiers;-)? Maybe IBM Watson Dialog contains a machine learning features? To be verified (my duty). About domain-specific I think that if domain is really "vertical", so a retrieval-based (or call I call"rule based pattern matching engine) like great Bruce Wilcox's ChatScript dialog flow language, could be a starting point for a possibile integration with some machine learning (a simple metalanguage for dialog flow design by linguistics, domain experts a deeplearning subsystem for inferences?) … some notes here:


Deep Spatio-Temporal Architectures and Learning for Protein Structure Prediction

Neural Information Processing Systems

Residue-residue contact prediction is a fundamental problem in protein structure prediction. Hower, despite considerable research efforts, contact prediction methods are still largely unreliable. Here we introduce a novel deep machine-learning architecture which consists of a multidimensional stack of learning modules. For contact prediction, the idea is implemented as a three-dimensional stack of Neural Networks NN k_{ij}, where i and j index the spatial coordinates of the contact map and k indexes ''time''. The temporal dimension is introduced to capture the fact that protein folding is not an instantaneous process, but rather a progressive refinement.


AI in 2018: Still more hype than reality, AI is nothing to be scared of yet

#artificialintelligence

Several weeks ago, I visited the website of Verizon Communications to check if a smartphone I wanted to buy on eBay was not stolen and would work on its network. A pop-up message asked me if I needed help, and even though I see chatbots often, I wondered if this one might possibly be human after it helped me and answered a few questions coherently.


Completion Reasoning Emulation for the Description Logic EL+

arXiv.org Artificial Intelligence

We present a new approach to integrating deep learning with knowledge-based systems that we believe shows promise. Our approach seeks to emulate reasoning structure, which can be inspected part-way through, rather than simply learning reasoner answers, which is typical in many of the black-box systems currently in use. We demonstrate that this idea is feasible by training a long short-term memory (LSTM) artificial neural network to learn EL+ reasoning patterns with two different data sets. We also show that this trained system is resistant to noise by corrupting a percentage of the test data and comparing the reasoner's and LSTM's predictions on corrupt data with correct answers.


Summarized Network Behavior Prediction

arXiv.org Machine Learning

This work studies the entity-wise topical behavior from massive network logs. Both the temporal and the spatial relationships of the behavior are explored with the learning architectures combing the recurrent neural network (RNN) and the convolutional neural network (CNN). To make the behavioral data appropriate for the spatial learning in CNN, several reduction steps are taken to form the topical metrics and place them homogeneously like pixels in the images. The experimental result shows both the temporal- and the spatial- gains when compared to a multilayer perceptron (MLP) network. A new learning framework called spatially connected convolutional networks (SCCN) is introduced to more efficiently predict the behavior.