Machine Learning

Deep Learning

20/06/2017 By debkr

I still haven’t had time to check out Google’s open source deep learning technologies but I know I need to get up to speed on this. They’ve now released Tensor2Tensor (T2T) to enable use of their machine intelligence library, TensorFlow, for training deep learning mobiles. Useful for translation, parsing, image captioning and lots of other great stuff, T2T comes with a library of research data sets and models.

Get T2T on GitHub: https://github.com/tensorflow/tensor2tensor

Original news article picked up from SDTimes: http://sdtimes.com/tensor2tensor-xebialabs-devops-syncsort-sdtimes-news-digest/

How We Find Out Things

01/03/2017 By debkr

How We Find Out Things is important. Intelligence is the ability to find things out about a situation or subject, and to apply what we have previously found out to a new or novel situations or subjects.

Machine Learning does this through methods of Computational Statistics. It is powerful, but we still need to be there applying our own Intelligence & Sense-making abilities to ensure the results are logical & consistent with what we would expect.

As Ramon y Cajal, Spanish biologist and father of modern neuroscience, noted in his excellent book of advice for novice scientific researchers (1999), the key stages of a scientific investigation are (1) observation, (2) experimentation, (3) working hypotheses, and (4) proof. [Read more…] about How We Find Out Things

MLND Project: Model Evaluation & Validation

10/02/2017 By debkr

3e Project: Model evaluation & validation

Project details

DESCRIPTION:

use Boston house price dataset to PREDICT selling price of a new/unseen home

PROCESS:

EXPLORE data > obtain important FEATURES & DESCRIPTIVE statistics abt data
Properly SPLIT dataset into TRAINING & TEST datasets
DETERMINE suitable PERFORMANCE METRIC for evaluating the problem
ANALYSE performance graphs for learning algorithm over varying TRAINING SET SIZES & with varying number of PARAMETERS
CHOOSE OPTIMAL MODEL that best generalises unseen data
TEST chosen optimal model on a NEW SAMPLE & COMPARE PREDICTED selling price to ACTUAL statistics

Vector Spaces

24/01/2017 By debkr

Thought on waking this morning:

Vector Spaces*

Had gone to bed asking: ‘OK so what should I do for our AI system? What’s next?’

Assume this answer means I should be studying Vector Spaces – and hence Linear Algebra – more fully for my delve into AI.

Makes sense given LA is such a strong math basis for all ML courses out there, also for AI generally.

But I can’t help wondering if this also relates somehow to my recent Systems Struggles. E.g. the vector space is an abstract mathematical way of representing a realworld physical geometric reality. Can this offer a parallel to Systems Modelling which might [Read more…] about Vector Spaces

Machine Learning Engineering

19/07/2016 By debkr

Dipping a toe in the Udacity pool:
I received an email from the UK/Europe Business Growth Manager of Udacity, looking for people with knowledge of/experiencein Machine Learning to lead and/or support face-to-face training sessions in Machine Learning in London. The email included the link to Udacity’s Nano Degree in Machine Learning.

While I’m not yet skilled or experienced enough at this time to lead a session, I felt really excited that I might have an opportunity to be a session supporter and assistant. That would be cool as I’d be exploring and networking in such an important growth area in tech skills, as well as supporting the growth of MOOCs here in London, which is something I feel really passionately about. [Read more…] about Machine Learning Engineering

Mathematics Mondays (revising some basics)

18/07/2016 By debkr

Here’s a quick revision and recap on some basic maths and linear algebra needed when studying Machine Learning. This will covers basics of calculus and linear algebra. A future post will revise some basic statistics, then I’ll go on todo some quick example calculations of standard deviations, derivatives and matrix determinants.

1. Linear Algebra – matrix multiplication:
Matrices can basically be though of as arrays of numbers, which might be a simple array (equating to a list in Python) which can be expressed as a simple one-row or one-column matrix:

, or they might be more complex matrices with multiple rows or columns: [Read more…] about Mathematics Mondays (revising some basics)

Mathematics Monday: Simple linear regression (revision)

13/06/2016 By debkr

Mathematical compass:
Someone recently suggested I make mathematics my compass (while I was busy diving down rabbit holes reading up on twistors and the like). Maybe they’re correct. I realised that if I was going to do that, I’d need to brush up on my map-reading skills first!

I’ve been so busy engrossed in learning Python I’ve sorely neglected the other streams of the training syllabus I set myself back in late February (is it really that long ago?). So I decided I need to dedicate at least one day a week to bringing my maths back up to speed.

I’m having to start almost back at square one. (Demoralising.) Calculus is a tad rusty to say the least: serious revision required. Did I really quit accountancy for this? Appears to be madness. Surely there’s some hidden logic in there somewhere, it just seems pretty elusive right now. Well, onwards and upwards anyway… [Read more…] about Mathematics Monday: Simple linear regression (revision)