Graphical Models Or Probabilistic And Unsupervised Learning. What Course Should I Take?

Jan 22 10 Comments

I am interested in pursuing a career as a quantitative researcher at a hedge fund/asset manager. I have to take either 'Graphical Models' or 'Probabilistic and Unsupervised Learning' as part of my masters degree in machine learning, the syllabus for the two courses are as follows:

Graphical Models:
-Bayesian Reasoning;
-Bayesian Networks;
-Directed and Undirected Graphical Models;
-Inference in Singly-Connected Graphs;
-Hidden Markov Models;
-Junction Tree Algorithm;
-Decision Making under uncertainty;
-Markov Decision Processes;
-Learning with Missing Data;
-Approximate Inference using Sampling;
-If time permits we will also cover some deterministic approximate inference

Probabilistic and Unsupervised Learning:
-Basics of Bayesian learning and regression;
-Latent variable models, including mixture models and factor models;
-The Expectation-Maximisation (EM) algorithm;
-Time series, including hidden Markov models and state-space models;
-Spectral learning;
-Graphical representations of probabilistic models;
-Belief propagation, junction trees and message passing;
-Model selection, hyperparameter optimisation and Gaussian-process regression

Thanks for any advice!

P.S. Please see the updated list of optional modules I have attached in picture format. I have to take 45-75 from the top group and then 45-15 credits from the bottom. Overall I must choose 90 credits

#HedgeFund #Quant #AI #MachineLearning

Graphical Models Or Probabilistic And Unsupervised Learning. What Course Should I Take?


Want to comment? LOG IN or SIGN UP
TOP 10 Comments
  • Ecolab / Data DataDork
    Graphical models sounds more like classical statistics, I believe that is more relevant to quant roles.
    Jan 22 1
    • JPMorgan Chase bGzW54
      Would you reccommend it over say NLP or Deep Learning? I have uploaded a screenshot of the list of available modules
      Jan 23
  • Uber / Data QajO06
    Sounds unsupervised one is more a fundamental course and might be helpful in interview/ jobs. The graphical models sounds more advanced and usually people in industry do not know this type of topic !!!
    Jan 22 1
    • JPMorgan Chase bGzW54
      Would you reccommend taking both, inplace of something like Deep Learning or NLP? Someone below suggested this as an option
      Jan 23
  • Indeed / Eng rainwater
    Honestly take both if you can
    Jan 22 3
    • Indeed / Eng rainwater
      What courses are you taking in total? That will make it easier for us to answer your question. I suggested both because they’re quite different but both subjects are used in quant finance. I haven’t heard Nlp being used directly in qf but it might be used in alternative data funds.
      Jan 22
    • JPMorgan Chase bGzW54
      Supervised learning is compulsory as well as statistics and data analysis. As for optional modules I had 6 to choose from with the two I mentioned being one of the compulsory choices. I was thinking of taking Intro to Deep Learning, Deep Reinforcement Learning, Statistical NLP, Numerical Optimisation and Advanced Topics in ML.

      I thought sentiment analysis was a big thing now hence my interest in NLP. Would you recommend replacing one of the above so I could take both Graphical Models and Probabilistic and Unsupervised Learning?
      Jan 22
  • Deloitte / Data Ceeker
    What school?
    I'd take graphical models and expand your understanding on your understanding on Bayesian principles and Decision Models/making .
    Also if you can take both do it
    NLP and deep learning are fringey
    Jan 22 1
    • JPMorgan Chase bGzW54
      This is for University College London in the U.K. it is their Computational Statistics and Machine Learning programme. There are options to take up to three finance related modules including Stochastic Methods in Finance I & II and a Financial Engineering course. From what I have read, Stochastic Processes used mainly in derivative pricing is no longer the main role of Quantitative Researchers (or atleast the skillset is not as sought after as say pre-08)

      Why do you say DL and NLP are fringey?
      Jan 22


    Real time salary information from verified employees