Blog Posts : /research/ideas/

Navigation

/ research
- / { ideas, talks, papers }
/ tech
- / { dl, talks, misc }

Tag search

Published on
September 13, 2022
Running MineRL within a ContainerMineRL
podman docker containers deeplearning minerl
Published on
July 29, 2020
TensorFlow 2.3 from source on Fedora 32Actually worked
Fedora Python TensorFlow
Published on
December 4, 2019
TensorFlow 2.1 from source on Fedora 30Actually worked
Fedora Python TensorFlow
Published on
December 11, 2017
NIPS 2017 ItineraryConference
Conference NeurIPS
Published on
October 2, 2017
Deep PocketsGoogle Reddit AMA (2017)
Google DNN
Published on
September 27, 2017
Failures, Hurdles & OpportunitiesGoogle Reddit AMA (2017)
Google DNN
Published on
September 18, 2017
Unsupervised LearningGoogle Reddit AMA (2017)
Unsupervised DNN
Published on
August 30, 2017
TensorFlow from source on Fedora 26Should be easy
Fedora Python TensorFlow
Published on
February 27, 2017
TensorFlow Estimator extra outputsSmuggling tensors out
Estimator TensorFlow
Published on
February 25, 2017
TensorFlow Estimator input_fnOld and New
Estimator CNN MNIST TensorFlow
Published on
December 27, 2016
Deep Learning FrameworksMNIST CNN - Rosetta Stone
MNIST DNN
Published on
December 10, 2016
Review of TensorFlow sugar coatingsCompare and Contrast
DNN TensorFlow
Published on
November 26, 2016
Layers on top of TensorFlowMNIST CNN - Rosetta Stone
MNIST DNN TensorFlow
Published on
August 28, 2016
Consciousness and the Social MachineThoughts
AGI PaperTakeaway Conciousness Attention-Schema Awareness Chatbots
Published on
November 11, 2015
TensorFlow InstallationPretty Smooth...
python TensorFlow Nvidia
Published on
November 10, 2015
Contributing to TensorFlow(we're not in <s>Kansas</s> GitHub any more)
theano python TensorFlow
Published on
October 27, 2015
A Small Theano Epiphany(may help someone)
theano python
Published on
October 23, 2015
Binary Classification vs Probability Regression{0,1} works better than (0,1)
DNN
Published on
October 19, 2015
Building Google's 1 Billion word corpusKaggle's version is identical
Corpus DNN
Published on
October 9, 2015
Theano and Blocks* in development mode
theano blocks git
Published on
August 31, 2015
Reinforcement Learning CoursePerformance-based Rewards (Musings)
DNN RL
Published on
August 27, 2015
GTC South Asia : NER PosterPoster
Poster NER theano blocks DNN
Published on
February 14, 2015
Batch Normalization - Accelerating Deep Network Training by Reducing Internal Covariate ShiftPaper Takeaway
PaperTakeaway DNN
Published on
November 17, 2014
TrailblazingMavericks Required
NLP PCA DNN
Published on
October 30, 2014
Evaluating entity linking with wikipedia - 2013Paper Takeaway
NLP NER NEL PaperTakeaway
Published on
October 14, 2014
Low Entropy Coding with Unsupervised Neural Networks - Harpur 1997Thesis Takeaway
PaperTakeaway DNN
Published on
October 13, 2014
GloVe - Global Vectors for Word Representation - Pennington, Socher, Manning 2014Paper Takeaway
NLP Word-Embeddings PaperTakeaway DNN
Published on
October 12, 2014
Learning word embeddings efficiently with noise-contrastive estimation - Mnih 2013Paper Takeaway
NLP Word-Embeddings PaperTakeaway DNN
Published on
October 7, 2014
Scalable Hierarchical Distributed Language Model - Mnih and Hinton 2009Paper Takeaway
NLP Word-Embeddings PaperTakeaway DNN
Published on
September 28, 2014
Neural Probabilistic Language Model - Bengio 2003Paper Takeaway
NLP Word-Embeddings PaperTakeaway DNN
Published on
September 17, 2014
Do Deep Nets Really Need to be Deep?Paper Takeaway
PaperTakeaway DNN
Published on
September 16, 2014
Combatting overfitting in TradingHyperparameters
AlgorithmicTrading MachineLearning
Published on
December 30, 2012
Neurons : Pyramids vs SheetsIs there really a pyramidal funnel from input to output?
Fundamentals DNN
Published on
December 23, 2012
Eye MovementsHow is the mind's eye built up in the brain?
Fundamentals
Published on
December 16, 2012
The Grandmother NeuronWhat to expect from a Grandmother Neuron
Fundamentals DNN
Published on
December 9, 2012
Sparse NetworksAre Neural Networks over-connected?
DNN
Published on
December 2, 2012
Real Neurons are Very ComplexHow realistic do we need our neural models to be?
DNN
Published on
November 25, 2012
How Much Domain Knowledge is Built-In?Not cheating at vision tasks
DNN
Published on
November 18, 2012
Taking Advantage of Paradigm ShiftsWhat's the most efficient way of getting to AI?
Philosophy Fundamentals
Published on
November 11, 2012
Where's the End of the Rainbow?And what should we expect to find there?
Philosophy Fundamentals
Published on
November 4, 2012
Learning without LabelsAvoiding local minima during childhood
Fundamentals DNN
Published on
October 28, 2012
Parallelism through TimeNeural Networks converge continuously
Parallelism DNN
Published on
October 21, 2012
Dropouts : 2^N networks for 2xN work?How 'dropout' relate to an idea from Genetic Algorithms
Fundamentals Dropout DNN
Published on
October 14, 2012
Unreliability & RobustnessWhat 'dropouts' mean for learning
Fundamentals Dropout DNN
Published on
October 7, 2012
Hierarchy and ConnectednessIs the brain pre-wired in a top-down hierarchy?
Philosophy Fundamentals DNN
Published on
September 30, 2012
How to Create a Map of ConceptsThere's no need to think too hard
Philosophy Fundamentals Concepts
Published on
September 23, 2012
Sleep and DreamsNarratives from noise
Philosophy Fundamentals Narratives
Published on
September 16, 2012
Miller's Law and its AI implicationsUseful chunks and other tradeoffs
Philosophy Fundamentals Concepts
Published on
September 9, 2012
Consciousness has a Narrative QualityWhat is your inner voice telling you?
Philosophy Fundamentals Narratives
Published on
September 2, 2012
Levels of ProcessingThe brain's scope may not be all bad news
Fundamentals DNN
Published on
August 26, 2012
Brain Homogeneity is Very EncouragingSimply replicating one good idea should scale up to AI (somehow)
Philosophy Fundamentals DNN
Published on
August 19, 2012
Bringing the Singularity CloserDesigning single purpose machines is relatively easier than generalized AI
Philosophy Fundamentals
Published on
August 12, 2012
There's no Shopping List for Artificial IntelligenceWhere's the Shopping List for the AI equivalent of the journey to Mars?
Philosophy Fundamentals

Blog Posts : /research/ideas/

Navigation

Tag search

Running MineRL within a ContainerMineRL

TensorFlow 2.3 from source on Fedora 32Actually worked

TensorFlow 2.1 from source on Fedora 30Actually worked

NIPS 2017 ItineraryConference

Deep PocketsGoogle Reddit AMA (2017)

Failures, Hurdles &amp; OpportunitiesGoogle Reddit AMA (2017)

Unsupervised LearningGoogle Reddit AMA (2017)

TensorFlow from source on Fedora 26Should be easy

TensorFlow Estimator extra outputsSmuggling tensors out

TensorFlow Estimator input_fnOld and New

Deep Learning FrameworksMNIST CNN - Rosetta Stone

Review of TensorFlow sugar coatingsCompare and Contrast

Layers on top of TensorFlowMNIST CNN - Rosetta Stone

Consciousness and the Social MachineThoughts

TensorFlow InstallationPretty Smooth...

Contributing to TensorFlow(we're not in <s>Kansas</s> GitHub any more)

A Small Theano Epiphany(may help someone)

Binary Classification vs Probability Regression{0,1} works better than (0,1)

Building Google's 1 Billion word corpusKaggle's version is identical

Theano and Blocks* in development mode

Reinforcement Learning CoursePerformance-based Rewards (Musings)

GTC South Asia : NER PosterPoster

Batch Normalization - Accelerating Deep Network Training by Reducing Internal Covariate ShiftPaper Takeaway

TrailblazingMavericks Required

Evaluating entity linking with wikipedia - 2013Paper Takeaway

Low Entropy Coding with Unsupervised Neural Networks - Harpur 1997Thesis Takeaway

GloVe - Global Vectors for Word Representation - Pennington, Socher, Manning 2014Paper Takeaway

Learning word embeddings efficiently with noise-contrastive estimation - Mnih 2013Paper Takeaway

Scalable Hierarchical Distributed Language Model - Mnih and Hinton 2009Paper Takeaway

Neural Probabilistic Language Model - Bengio 2003Paper Takeaway

Do Deep Nets Really Need to be Deep?Paper Takeaway

Combatting overfitting in TradingHyperparameters

Neurons : Pyramids vs SheetsIs there really a pyramidal funnel from input to output?

Eye MovementsHow is the mind's eye built up in the brain?

The Grandmother NeuronWhat to expect from a Grandmother Neuron

Sparse NetworksAre Neural Networks over-connected?

Real Neurons are Very ComplexHow realistic do we need our neural models to be?

How Much Domain Knowledge is Built-In?Not cheating at vision tasks

Taking Advantage of Paradigm ShiftsWhat's the most efficient way of getting to AI?

Where's the End of the Rainbow?And what should we expect to find there?

Learning without LabelsAvoiding local minima during childhood

Parallelism through TimeNeural Networks converge continuously

Dropouts : 2^N networks for 2xN work?How 'dropout' relate to an idea from Genetic Algorithms

Unreliability & RobustnessWhat 'dropouts' mean for learning

Hierarchy and ConnectednessIs the brain pre-wired in a top-down hierarchy?

How to Create a Map of ConceptsThere's no need to think too hard

Sleep and DreamsNarratives from noise

Miller's Law and its AI implicationsUseful chunks and other tradeoffs

Consciousness has a Narrative QualityWhat is your inner voice telling you?

Levels of ProcessingThe brain's scope may not be all bad news

Brain Homogeneity is Very EncouragingSimply replicating one good idea should scale up to AI (somehow)

Bringing the Singularity CloserDesigning single purpose machines is relatively easier than generalized AI

There's no Shopping List for Artificial IntelligenceWhere's the Shopping List for the AI equivalent of the journey to Mars?

Failures, Hurdles & OpportunitiesGoogle Reddit AMA (2017)