Видео

In this session, we spend a bit of time learning about Quarto, allowing you to create good-looking reports that weave together code, text, images and more in a neat, easily-distributable format. We also have a look at using a Python script to interact with the command line and produce multiple outputs from a single Quarto file using different parameters. We will then briefly dive into the world...

We kick off this session with a recap of the key Streamlit commands (though we'd recommend going through session 7B if you're new to Streamlit: ruclips.net/video/iOfzfmuAr_Y/видео.html) before a bit of a blast from the past - returning to our discrete event simulation (DES) code of a walk-in clinic from module 2 and turning it into an app. We then move on to sidebars, run buttons, multi-page ap...

Today we're going to start looking at building web apps entirely in Python using the fantastic Streamlit framework. HSMA Streamlit Book: bergam0t.github.io/streamlit_book/ Slides: docs.google.com/presentation/d/1F3Fnn8OF-m02NClaHICvM5LDWrkV20np-bhnBfpKOvA/edit?usp=sharing Github Repository: github.com/hsma-programme/h6_7b_web_apps_1 HSMA Website: sites.google.com/nihr.ac.uk/hsma/home

Today we're going to be looking at the crucial importance of version control, and how we can use Git - a distributed version control system - to allow us to this. We'll also look at how we can use Git interaction with GitHub to publish and work with repositories hosted online. Git is both amazing and horrible. If you feel that this all seems odd today, and convoluted, you're not wrong. But it I...

In this session, we look at how we can use the excellent Python package MESA to help us build our own Agent Based Simulations, and we'll revisit our old friend Object Oriented Programming to help us do that. Slides: docs.google.com/presentation/d/14s13gg3XI_MzG47nq73R1VPpXkLIhG7NfsrgF_uWWzQ/edit?usp=sharing Github Repository: github.com/hsma-programme/h6_6c_abs_part_2 HSMA Website: sites.google...

In this session, we take a look at System Dynamics as a way of understanding whether there are inherent issues in the way our systems are structured. We use the online InsightMaker (insightmaker.com/) tool to create a quantitative system dynamics model. Slides: docs.google.com/presentation/d/12NHXp_uou7ayDAcTJNgXwr0_O3rQdrIRmq6wkJG_2tA/edit?usp=sharing Repository: github.com/hsma-programme/h6_6...

In this session, we begin looking at Agent Based Simulation - a simulation modelling approach that allows us to capture individual level behaviours of actors in our system, and observe the emergent dynamics. We also take a look at Cellular Automata, and head to the Edge of Chaos. Slides: docs.google.com/presentation/d/1of8eQAE79IFCJawvtNu3-ooesLtRzO4dXp7zWjL4jek/edit?usp=sharing Repository: git...

In this session, HSMAs worked in groups to tackle a question of their choosing on some data of their choosing using Natural Language Processing techniques, undertaking miniature projects and practicing those all-important skills needed to tackle their own projects in Phase 2. The recording contains the brief for the day, as well as 10 presentations of project plans after the initial hour, and 1...

In this session we don our pirate outfits and sail to the Caribbean as we look for the secret of Sentiment Analysis and Text Classification. We'll fight dastardly pirates as well as a pirate that's already dead but just won't accept it. Slides: docs.google.com/presentation/d/1k2gQkA6Qa0TKm47Hh6tM7heTpYh6w2wF27ImQC_BeXk/edit?usp=sharing Repository: github.com/hsma-programme/h6_5c_sentiment_analy...

In this session we'll continue our journey into Natural Language Processing. Have you got named entities? Do you need them recognised? Then have we got the session for you! Our session on Named Entity Recognition will talk all about named entities, and recognition of them. Along the way we'll stop by Sesame Street and Fraggle Rock, talk about Transformers, Parsing, Grammar, POS Tagging and more...

In this session we'll begin Module 5 and our journey into Natural Language Processing. We'll look at some core concepts important for processing text data, and learn a few little neat things we can do with our data along the way, including wordclouds and a method for assessing the similarity of different bits of text. Slides: docs.google.com/presentation/d/15NWc3d84XusMmTHQdpIgfvD6LdUGfZwk7cLyL...

Unfortunately the first 5 minutes or so of the lecture was not recorded Covering a range of ways to improve your model's performance, including: - Missing Data Imputation with SimpleImputer and IterativeImputer - Feature Selection with SequentialFeatureSelector (forward and backward selection) and SelectFromModel (feature importance selection with model coefficients or mean decrease in impurity...

In this session we take a look at Reinforcement Learning in a session that will be very different to any you've experienced thus far. Note that we'd recommend not looking at the slides until after the first time the reinforcement learning game is played manually. Slides: docs.google.com/presentation/d/1Sok1fWn_y15XLRLQoOq2pKiTX6TG3LDgNkSKhFq6us0/edit?usp=sharing App Link: bergam0t.github.io/Rei...

In this session we take a look at synthetic data - how to create our own fake but realistic data when we want to augment an underrepresented class, or just use the data instead of our real data. Slides: docs.google.com/presentation/d/1psVyGCwFDAtW0v7QAUaWXseUTV7W-xgRPDgS7rtjpJU/edit?usp=sharing Github Repository: github.com/hsma-programme/h6_4i_synthetic_data HSMA Website: sites.google.com/nihr...

(HSMA 6 Day 12) 4G Part 5/6 - Explainable AI with SHAP

(HSMA 6 Day 12) 4G Part 6/6 - Prediction Uncertainty

(HSMA 6 Day 13) 4F Neural Networks

(HSMA 6 Day 12) HSMA Assessment Update

(HSMA 6 Day 13) 4G Part 4/6 - Partial Dependence Plots and Individual Conditional Expectation Plots

(HSMA 6 Day 13) 4G Part 3/6 - Feature Importance with MDI PFI

(HSMA 6 Day 13) 4G Part 2/6 - Feature Importance in Logistic Regression & Odds/Log Odds/Probability

(HSMA 6 Day 13) 4G Part 1/6 - An Introduction to Explainable AI Correlation vs Causality

(HSMA 6 Day 12) 4E Part 1/2 - Boosting Trees for Classification Problems

(HSMA 6 Day 12) 4E Part 2/2 - Predicting Numeric Values with Tree-based Algorithms OneHot encoding

(HSMA 6 Day 12) 4D Part 2/2 - Random Forests, F1 score Confusion Matrices

(HSMA 6 Day 12) 4D Part 1/2 - Decision Trees

(HSMA 6 Day 11) 4C - Ethics in AI

(HSMA 6 Day 11) 4B - Logistic Regression - Who Would Survive the Titanic?

(HSMA 6 Day 10) 4A - Introduction to AI and Machine Learning