CSIR Machine Learning Reading Group 


Reading group meetings are held
Every Thursday, at 12pm, in the MIAS Boardroom (Building 17A, first floor).
All welcome - no experience required!

Please join the mailing list: groups.google.com/d/forum/csir-mds-ml
Note: you need to ensure that you are notified by email by the group!

MLRG Slack channel (MDS only): https://csirmds.slack.com/messages/mlrg/


This group is managed by:
  • Benjamin Rosman (brosman@csir.co.za).
  • Michael Burke
  • Vukosi Marivate
  • Nyalleng Moorosi
  • Christiaan van der Walt
Please contact us with any comments. 



Potential future themes (please send us any suggestions):
  • Support Vector Machines (and other kernel methods)
  • Decision trees/forests
  • Probabilistic graphical models
  • Reinforcement learning
  • Deep learning
  • Optimization techniques
  • Sampling
  • Handling temporal/sequential data

Current rota:
  • Theme 9: Ensemble learning
  • Theme 8: Reinforcement learning
  • Theme 7: Probabilistic graphical models
  • Theme 6: Deep learning         
  • Theme 5: Optimisation          
  • Theme 4: Visualisation
    • Introduction to visualisation and analytics - Vukosi Marivate (27 Oct)
    • Visualisation through mapping - Bolelang Sibola (3 Nov)
    • No meeting (10 Nov)
    • No meeting (17 Nov)
    • Lying with visualisation - Vukosi Marivate (24 Nov)
  • Theme 3: Dimension reduction
    • No meeting (15 Sept)
    • No meeting (22 Sept)
    • PCA and PPCA - Ari Ramkilowan (29 Sept)
    • Kernel PCA + some extensions - Ashley Kleinhans (6 Oct)
    • Sparse dictionary learning - Luke Darlow (13 Oct)
    • Locality-sensitive hashing
       - Tulani Mzayidume (20 Oct) - slides

  • Theme 2: Intro to classification
    • Linear classifiers - Fisher's linear discriminant, logistic regression, Naive Bayes classifier - Ofentswe Lebogo (11 August) - slides
    • Decision trees - Leilanie Uys (18 August) - slides
    • No meeting (25 August)
    • K-nearest neighbours
       - Vukosi Marivate (1 Sept) - slides, github repo
    • Support vector machines
       Christiaan van der Walt (8 Sept)

  • Theme 1: Intro to regression
    • Introduction to machine learning - Benjamin Rosman (14 July) - slides
    • Linear regression - Robert Berman (21 July) - slides
    • Kernels and regularization - Christiaan van der Walt (28 July) - slides
    • Bayesian linear regression - Michael Burke (4 August) - (slides+notebooks)

  • Theme 0: Bayesian rota selection
 Topic votes



2015 Rota: 

  • Theme 5: Neural Networks and Deep Learning (led by B Rosman)
    • A Gentle Introduction to Neural Networks and Deep Learning - Benjamin Rosman (29 Oct) - slides
    • A Hitchhiker's Guide to Neural Networks - Todani Luvhengo (5 Nov) - slides
    • Convolutional Neural Networks - Beatrice van Eden (12 Nov) - slides
  • No meeting (22 Oct)
  • Theme 4: Evaluation of Learning (led by V Marivate)
    • The Philosophy and History of Evaluation - Vukosi Marivate (1 Oct) - slides
    • Evaluation for Supervised Learning - Ofentswe Lebogo (8 Oct) - slides
    • Evaluation for Unsupervised Learning - Patrick Monamo (15 Oct) - slides
  • No meeting - Heritage Day (24 Sept)
  • Theme 3: Support Vector Machines (led by C van der Walt)
    • Overview of Support Vector Machine Classification - Christiaan van der Walt (10 Sept) - slides
    • Support Vector Machine Kernels and Hyperparameter Optimisation - Melvin Diale and Nthabiseng Mokoena - slides (17 Sept)
  • Catch-up (3 Sept)
    • Discussion of the direction of the group, and any interesting ML problems
  • Theme 2: Dimension Reduction (led by M Burke)
    • Introduction to dimension reduction - Michael Burke (6 Aug) - slides and code
      • Covering the curse of dimensionality, properties of features, and worked examples illustrating these
    • Unsupervised linear dimension reduction - Ardhisha Pancham (13 Aug) - slides and code
      • Covering PCA and NMF
    • Supervised linear dimension reduction - Todani Luvhengo (20 Aug) - slides and code
      • Covering LDA and Canonical variates
      • Python links:
        • Anaconda http://continuum.io/downloads
        • Python packages
          • http://pandas.pydata.org/ (pandas)
          • http://www.scipy.org/ (scipy)
          • http://www.numpy.org/ (numpy)
          • http://scikit-learn.org/ (sklearn)
    • Non-linear dimension reduction - Yongama Feni (27 Aug) - slides and code
      • Covering Isomap and LLE
  • Theme 1: Introduction (led by B Rosman)
    • General Introduction to Machine Learning - Benjamin Rosman (9 July) - slides
    • Introduction to Supervised Learning - Ashley Kleinhans (16 July) - slides
    • Introduction to Unsupervised Learning - Beatrice van Eden (23 July) - slides
    • Tools to get Started - Robert Berman (30 July) - slides



For previous rotas and slide links go to: Annual archive 


Subpages (1): Annual archive
Ċ
Benjamin Rosman,
Jul 14, 2016, 6:46 AM
Ċ
Benjamin Rosman,
Sep 10, 2015, 6:26 AM
ć
ConvNet.pptx
(12490k)
Beatrice van Eden,
Nov 15, 2015, 10:05 PM
ć
Benjamin Rosman,
Aug 22, 2016, 6:34 AM
ċ
DimensionReductionTalk.zip
(182k)
Benjamin Rosman,
Aug 6, 2015, 6:43 AM
Ċ
ESL.pdf
(409k)
Beatrice van Eden,
Oct 8, 2015, 3:02 AM
Ċ
Benjamin Rosman,
Oct 1, 2015, 3:47 AM
Ċ
Michael Burke,
Aug 11, 2016, 5:12 AM
Ċ
Benjamin Rosman,
Jul 22, 2016, 3:12 AM
ċ
Michael Burke,
Oct 15, 2015, 3:13 AM
ċ
Machine Learning-Dimension Reduction.rar
(1847k)
Benjamin Rosman,
Aug 27, 2015, 4:49 AM
ċ
Notebooks_data.zip
(54k)
Michael Burke,
Jul 13, 2016, 12:23 PM
Ċ
Benjamin Rosman,
Sep 17, 2015, 4:11 AM
Ċ
Ashley Kleinhans,
Jul 16, 2015, 2:54 AM
Ċ
Benjamin Rosman,
Aug 5, 2016, 2:41 AM
Ċ
Robert Berman,
Jul 31, 2015, 7:24 AM
ć
Benjamin Rosman,
Oct 20, 2016, 6:15 AM
Ċ
Beatrice van Eden,
Jul 27, 2015, 4:50 AM
ċ
UnsupervisedLinearDimensionReduction.zip
(3034k)
Beatrice van Eden,
Aug 14, 2015, 1:42 AM
ċ
bayes_04_08.zip
(469k)
Michael Burke,
Aug 4, 2016, 4:26 AM
Ċ
Benjamin Rosman,
Oct 29, 2015, 4:47 AM
Ċ
Benjamin Rosman,
Nov 5, 2015, 2:29 AM
ċ
supervised_linear_dimensionality_reduction.zip
(281k)
Beatrice van Eden,
Aug 24, 2015, 6:06 AM