Half-Day Training | Deep Learning | Machine Learning | Beginner-Intermediate

We will use OpenAI gym to try our RL algorithms. OpenAI is a non profit organization that want committed to open source all their research on Artificial Intelligence. To foster innovation OpenAI created a virtual environment, OpenAi gym, where it’s easy to test Reinforcement Learning algorithms. In particular, we will start with some popular techniques like Multi-Armed Bandit, going thought Markov Decision Processes and Dynamic Programming. We then will also explore other RL frameworks and more complex concepts like Policy gradients methods and Deep Reinforcement learning, which recently changed the field of Reinforcement Learning. In particular, we will see Actor-Critic models and Proximal Policy Optimizations that allowed OpenAI to beat some of the best Dota players. We will also provide the necessary Deep Learning concepts for the course…more details

Full-Day Training | NLP | Deep Learning | Intermediate-Advanced

Being specialized in domains like computer vision and natural language processing is no longer a luxury but a necessity which is expected of any datascientist in today’s fast-paced world! With a hands-on and interactive approach, we will understand essential concepts in NLP along with extensive hands-on examples to master state-of-the-art tools, techniques and methodologies for actually applying NLP to solve real- world problems. We will leverage machine learning, deep learning and deep transfer learning to learn and solve popular tasks using NLP including NER, Classification, Recommendation \ Information Retrieval, Summarization, Classification, Language Translation, Q&A and Topic Models…more details

Half-Day Training | Deep Learning | Machine Learning | Beginner-Intermediate

We will use OpenAI gym to try our RL algorithms. OpenAI is a non profit organization that want committed to open source all their research on Artificial Intelligence. To foster innovation OpenAI created a virtual environment, OpenAi gym, where it’s easy to test Reinforcement Learning algorithms. In particular, we will start with some popular techniques like Multi-Armed Bandit, going thought Markov Decision Processes and Dynamic Programming. We then will also explore other RL frameworks and more complex concepts like Policy gradients methods and Deep Reinforcement learning, which recently changed the field of Reinforcement Learning. In particular, we will see Actor-Critic models and Proximal Policy Optimizations that allowed OpenAI to beat some of the best Dota players. We will also provide the necessary Deep Learning concepts for the course…more details

Half-Day Training | Data Visualization | | Intermediate

In this session, you’ll learn how to turn any matplotlib plot into an animated gif with just two extra lines of code (well… three, if you count “import gif”)!..more details

Full-Day Training | Machine Learning | Beginner-Intermediate

This session is a hands-on introduction to Machine Learning in Python with scikit-learn. You will learn to build and evaluate predictive models on tabular data using the main tools of the Python data-science stack (Jupyter, numpy, pandas, matplotlib and … scikit-learn)…more details

Tutorial | Data Visualization | All Levels

Coming Soon!

Half-Day Training | Machine Learning | R-programming | Intermediate-Advanced

Temporal data requires special care to model as it violates several principles of standard machine learning models. R has long had top-of-the-line forecasting tools, though recently new ones have been developed which greatly ease working with time series data. We use the tsibble package for manipulating time series data, feasts for visualization , and fable for building forecasting models such as ETS and ARIMA….more details

Half-Day Training | MLOps & Data Engineering | Beginner-Intermediate

In this workshop, we show how to maintain data science productivity as well as collaborate effectively and deliver value continuously and seamlessly. We demonstrate and guide the participants through CI/CD practices for machine learning and a new pattern of working that avoids most of the pitfalls of the typical approach.

Participants will learn how to utilize new patterns of repeatable continuous model development to collaborate effectively and deliver value continuously and seamlessly in industrial data science projects using Continuous Integration (CI) and Continuous Delivery (CD) practices...more details

Half-Day Training | R-Programming | Beginner

Should a data scientist use R or should they use Python? The answer to this rather delicate question is, of course, they should know a bit each and use the most appropriate language for the task in hand. In this tutorial, we’ll take participants through the R tidyverse, one of the (many!) areas that R shines. The tidyverse is essential for any data scientist who deals with data on a day-to-day basis. By focusing on small key tasks, the tidyverse suite of packages removes the pain of data manipulation. We’ll cover some of the core features of the tidyverse, such as dplyr (the workhorse of the tidyverse), string manipulation, graphics and the concept of tidy data…more details

Half-Day Training | Machine Learning | R-programming | Beginner

Linear regression is the foundation of supervised learning, though it has its limits. During this workshop we extend regression using penalization for automated variable selection and increased flexibility. We then introduce trees, and in particular boosted trees, via xgboost to get incredibly powerful predictions. We will go over some of the theory and also practical considerations such as hyperparameters…more details

Workshop | NLP | Research Frontiers | Intermediate-Advanced

The speaker will review case studies from real-world projects that built AI systems using Natural Language Processing (NLP) in healthcare. These case studies cover projects that deployed automated patient risk prediction, automated diagnosis, clinical guidelines, and revenue cycle optimization. He will also cover why and how NLP was used, what deep learning models and libraries were used, and what was achieved. Key takeaways for attendees will include important considerations for NLP projects including how to build domain-specific healthcare models and using NLP as part of larger and scalable machine learning and deep learning pipelines in distributed environment…more details

Workshops | Deep Learning | Beginner-Intermediate

Learn the basics of building a PyTorch model using a structured, incremental and from first principles approach. Find out why PyTorch is the fastest growing Deep Learning framework and how to make use of its capabilities: autograd, dynamic computation graph, model classes, data loaders and more.
The main goal of this session is to show you how PyTorch works: we will start with a simple and familiar example in Numpy and “torch” it! At the end of it, you should be able to understand PyTorch’s key components and how to assemble them together into a working model.
We will use Google Colab and work our way together into building a complete model in PyTorch. You should be comfortable using Jupyter notebooks, Numpy and, preferably, object oriented programming…more details

Half-Day Training | Machine Learning | Intermediate

How could a violation of the triangle inequality theorem in mathematics lead to a cure for cancer? How can a mathematical concept from the 18th century be used to estimate the mass density of galaxies across the Universe? How could a marketing segmentation algorithm protect astronauts traveling to Mars from certain death? How does a F1 race from the 1950’s inspire one of the greatest machine learning use cases for the Internet of Things? This workshop will answer these questions, and more, by presenting several examples of typical algorithms that were adopted for specific use cases or application domains, then showing how each one can be adapted to an atypical (often mind-bending) use case, producing significantly surprising results in some other domain. These exercises serve to demonstrate how data scientists can create even more value, beyond that which is expected, from our data sets and our algorithmic talents…more details

Workshop | ML for Programmers | Open-source | Intermediate

Pomegranate is a Python package for probabilistic modeling that emphasizes both ease of use and speed. In keeping with the first emphasis, pomegranate has a convenient modular API for building complex models out of simple components and a simple sklearn-like API for using them. In keeping with the second emphasis, the computationally intensive parts of pomegranate are written in Cython and all models support multithreaded parallelism and out-of-core computations. In this talk I will give an overview of the features in pomegranate, including missing value support and the new data generators, and demonstrate how they can yield more accurate models. I will also demonstrate how one can use existing components to build neural probabilistic models, such as neural HMMs, using your favorite neural network package…more details

Half-Day Training | NLP | Beginner

Coming Soon!

Half-Day Training | Open-source | Beginner

This tutorial will introduce you to the wonderful world of Bayesian data science through the lens of probabilistic programming in Python. In the first half of the tutorial, we will introduce the key concepts of probability distributions via hacker statistics, hands-on simulation, and telling stories of the data-generation processes. We will also cover the basics of joint and conditional probability, Bayes’ rule, and Bayesian inference, all through hands-on coding and real-world examples. In the second half of the tutorial, we will use a series of models to build your familiarity with PyMC3, showcasing how to perform the foundational inference tasks of parameter estimation, group comparison (for example, A/B tests and hypothesis testing), and arbitrary curve regression. By the end of this tutorial, you will be equipped with a solid grounding in Bayesian inference, able to write arbitrary models, and have experienced basic model checking workflow…more details

Half-Day Training | Data Visualization | MLOps & Data Engineering | Beginner

Learn the art and science, the theory and practical hands-on tactics of creating compelling communications experiences based on surveys, scientific research, public or proprietary data sources. In this talk, you will learn about some of the best practices for data storytelling and visualization and walk away with a new way of thinking and some practical tips you can put to work the moment the session is done…more details

Half-Day Training | Machine Learning | Intermediate

Unsupervised learning models (including analysis of correlations, clusters, and associations in data) converge more readily to a useful solution if we start with good model parameterizations. Feature engineering is key, but selection of features often becomes guesswork. Similarly, in supervised machine learning, the choice of features in labeled data to use in training may still seem arbitrary. So, how does model-building start and move towards an optimal solution? This challenge is known as the cold-start problem! The solution to the problem is easy (sort of): We start with a guess, a totally random guess! That sounds so random, and so wrong! But there is an orderly and productive way forward from such a start, which we will describe in this workshop. We will present several machine learning modeling examples, suggested solutions to their cold-start challenges, and related concepts, including the objective function, genetic algorithms, backpropagation, gradient descent, and meta-learning…more details

Half-Day Training | Deep Learning | Advanced

Although supervised learning has dominated industry machine learning implementations, unsupervised and semi-supervised methods have started to be practically applied to real world problems (outside of playing video games). Generative Adversarial Networks (GANs) are being utilized to augment data and generate dialogue, and Reinforcement Learning (RL) is helping people plan marketing campaigns and control robots.

In this training, you will develop a theoretical understanding of these and other related state-of-the-art AI methods along with the hands-on skills needed to train and utilize them. You will implement a variety of models in TensorFlow for tasks including object recognition, image generation and robotics…more details

Half-Day Training | MLOps & Data Engineering | R-programming | Intermediate

Modern machine learning is mostly brute forcing through a multitude of hyperparameters. We look at new tools for building and tuning models in R, some so new they are only available on GitHub. We use xgboost and glmnet models (learned in Part I) as motivation for learning how to split data and conduct cross-validation with rsample, perform feature engineering with recipes, build model specifications with parsnip, tune over hyperparameters with dials and tune, evaluate performance with yardstick and put it all together with workflows…more details

Tutorial | Machine Learning

Coming Soon!

Tutorial | Research Frontiers | Machine Learning | Intermediate

Many real-world phenomena such as epidemics spreading, traffic flow, team sports, and climate variation involve complex spatial patterns evolving through time. Our ability to learn and mine from large-scale spatiotemporal data is critical to real-time decision making in various science and engineering fields. However, existing machine learning methods are still insufficient for large-scale spatiotemporal data due to the presence of non-linear, non-Euclidean, multi-resolution, high-dimensional, and complicated physical characteristics. The goal of this tutorial is to (1) provide an overview of the nature of spatiotemporal data and its relevance to the data mining community (2) survey recent development in machine learning to address the challenges specific to spatiotemporal data, and (3) identify the open problems and future directions. We believe this is an emerging and high-impact topic in machine learning, which will attract both academics and practitioners in data science as well as domain scientists…more details

Bootcamp | Kickstarter | Open-source | Beginner

By completing this workshop, you will develop an understanding of relational models of data, how SQL is used to retrieve that data, and how to join tables, aggregate information, and answer data science questions. You will also become familiar with many of the common types of SQL databases, how to access information in a database from the command line, and how to integrate database access from within Python.

Lesson 1: Relational Databases and Foundational SQL
Familiarize yourself with relational databases and the SQL syntax necessary to retrieve information from tables in a database. At the end of this lesson, you will be able to comfortably explore a database and retrieve filter, and sort information from a table…more details

Half-Day Training | Kick-starter | Open-source | Intermediate

In this training, you will learn how to accelerate your data analyses using the Python language and Pandas, a library specifically designed for tabular data analysis. We start by learning the core Pandas data structures, the Series and DataFrame. From these foundations, we will learn to use the split-apply-combine paradigm for grouped computations, manipulate time series, and perform advanced joins between datasets. Specifically, loading, filtering, grouping, and transforming data. Having completed this workshop, you will understand the fundamentals and advanced features of Pandas, be aware of common pitfalls, and be ready to perform your own analyses…more details

Half-Day Training | Deep Learning | Machine Learning | Intermediate-Advanced

Welcome to deep learning building blocks. This is an intermediate tutorial on deep learning that focuses on how to design neural networks for various data types.…more details

Half-Day Training | Quant Finance | Machine Learning | Intermediate

The rapid progress in machine learning (ML) and the massive increase in data availability has enabled novel approaches to quantitative investment and increased the demand for the application of data science to develop discretionary and automated trading strategies.

This workshop covers popular ML use cases for the investment industry. In particular, it focuses on how ML fits into the workflow of developing a trading strategy, from the engineering of financial features to the development of an ML model that generates tradable signals, the backtesting of a trading strategy that acts on these signals and the evaluation of its performance.

We’ll use common Python data science and ML libraries as well as Zipline, Pyfolio, and Alphalens. The code examples will be presented using jupyter notebooks and are based on the second edition of my book ‘Machine Learning for Algorithmic Trading’...more details

Tutorial | Machine Learning | MLOps & Data Engineering | Intermediate

ONNX Runtime is the inference engine for inferencing ML models in different HW environments. Applications of AI are everywhere. This requires ML models that are trained in the cloud to execute on small devices with low power, low compute, and low memory. Such devices are typically used in IoT scenarios. The data captured by these devices are processed before sending the telemetry to the cloud for further actions on the business application. ONNX Runtime has made enhancements to enable execution of ML models in these edge devices to power AI on the edge applications. This session will walk through the workflow to train an image classification model, package in container and deploy to IoT device…more details

Half-Day Training | MLOps & Data Engineering | R-programming | Intermediate-Advanced

After we build various machine learning models we need to make them accessible to others. We use modeldb so we can store models in a database and plumber to expose our model as a REST API that can be hosted in a Docker container…more details

Workshop | Machine Learning | Open-source | Intermediate-Advanced

In this workshop, we will work hands-on using XGBoost with real-world data sets to demonstrate how to approach data sets with the twin goals of prediction and understanding in a manner such that improvements in one area yield improvements in the other. Using modern tooling such as Individual Conditional Expectation (ICE) plots and SHAP, as well as a sense of curiosity, we will extract powerful insights that could not be gained from simpler methods. In particular, attention will be placed on how to approach a data set with the goal of understanding as well as prediction…..more details

Workshops  | NLP | Machine Learning | Intermediate

For BlueVine, and indeed for any Fintech company, figuring out the client’s industry is a critical factor in making precise financial decisions. Traditional sources are invariably pricey, inaccurate and unavailable, and as such leave an opening for an ML based solution. We met that challenge building a service that predicts the industry using the business’s publicly available web data. By employing the latest innovations in NLP (BERT) and some of the most powerful scraping and deployment tools available (Scrapy and Amazon SageMaker) we were able to dramatically surpass the performance achieved by any other such tool in the space. This presentation will cover the entire development pipeline hands-on: Crowdsourcing a tagged sample, building a smart and scalable web scraper, prepping and feeding the resulting raw data into BERT, fine tuning the model and finally deploying it as a cloud based service behind an API. Both model training and deployment will be through Amazon SageMaker…more details

Workshop | Deep Learning | Data Visualization

Coming Soon!

Tutorial | NLP | MLOps & Data Engineering | Beginner

Advances in Natural Language Processing (NLP) over the last year have been changing the way we work with text-based data. We have seen the release of Google’s BERT, OpenAI’s GPT-2, and Hugging Face’s Transformers library that has helped many use these kinds of models. But it can still be challenging for many users to get started and apply them to their own datasets.
To address this challenge, Novetta has open sourced AdaptNLP, an intuitive framework that lowers the barrier to entry to use these advanced capabilities. This high-level framework enables users to use fine-tune pre-trained language models for text classification, question answering, entity extraction, and part-of-speech tagging. This tutorial will provide quick and easy access to a variety of embedding schemes for downstream use. The ability to stand each of these tasks up as a service for easy integration into existing workflows and applications becomes fast and straightforward....more details

Workshop | Machine Learning | | Intermediate

I will present an overview of causal inference techniques that are a good addition to the toolbox of any data scientist, especially in certain circumstances where experimentation is limited. Use of these techniques can provide additional value from historical data as well to understand drivers of key metrics and other valuable insights. The session will be practical focused with both theory and how to perform techniques in R. The end of the session will close with recent advances from combining machine learning with causal inference techniques to do things such as speed up AB testing…more details

Tutorial | Machine Learning | Open-source | Beginer-Intermediate

Faces are a fundamental piece of photography, and building applications around them has never been easier with open-source libraries and pre-trained models.

In this tutorial, we’ll help you understand some of the computer vision and machine learning techniques behind these applications. Then, we’ll use this knowledge to develop our own prototypes to tackle tasks such as face detection (e.g. digital cameras), recognition (e.g. Facebook Photos), classification (e.g. identifying emotions), manipulation (e.g. Snapchat filters), and more...more details

Half-Day Training | Kickstarter | Beginner

Programming is a great way to get practical insights about math theoretical concepts. The goal of this session is to show you that you can start learning the math needed for machine learning and data science using code. You’ll learn about scalars, vectors, matrices and tensors, and see how to use linear algebra on your data. Don’t worry if you don’t have a math background, we’ll explain the mathematical notations and conventions. At the end of the session, you’ll know how to use the norm of vectors, and apply the concept to model regularization (Hands-on Session 1.) and cost functions. You’ll also see how to manipulate matrices and tensors with Numpy, and learn to use them to store and classify images (Hands-on Session 2).…more details

Tutorial | MLOps & Data Engineering | Intermediate-Advanced

The workshop is aimed at data scientists and data engineers who are interested in building a production-ready data pipelines

We will go into:

1. The challenges associated with creating ML models that are deployable
2. Software engineering principles that should be applied to ML code to make it easier to deploy in the production environment
3. How you can use an open-source Python library, called [Kedro](https://github.com/quantumblacklabs/kedro), to enhance your exploratory data analysis workflow as well as their transition to production-ready code.…more details
   

Tutorial | Deep Learning | All Levels

Transformer has been around for a while now, and has proven to be one of the most interesting models of modern deep learning.

It turns out that Transformer models have proven to be domain agnostic. It means that, that despite of its initial application for seq2seq NLP tasks with 1D sequences of text, the 1D transformer input can be of any form. Namely, a 2D image unrolled into long 1D sequence of pixels can be understood with notion of its 2D image characteristics involving object appearance, category, or even predicting next image appearance in very long sequences.

Recent research show that transformer originated architectures for computer vision often tends to be simpler and provide performance at worst on pair with modern architectures such as RCNNs used for the computer vision tasks. Presentation is going to discuss recent research in area of transformer applications in new domains..more details

Talk | MLOps & Data Engineering | Intermediate

We need to collect ever larger amounts of data! Why? Because we might find patterns of behaviour in our macro world –people, environment, upper atmosphere and deep oceans- and in our micro world – disease vectors, the heart of the matter (currently quarks) and so on. Once we have filled the big data silos, we will not only find patterns of (hidden) behaviour but we will be able to make machines learn the behaviour and mimic the operations performed by clever humans in the macro- and micro world. Then we have a disaster created, say, by greed, and all the systems that have discovered and learnt the behaviour of the ‘market’ failed to work costing trillions of dollars, yes and euros: We find that there was yet another pattern of behaviour we had failed to discover and learn – the market sentiment was what we had missed so the claim goes. Oh dear. A race is on to collect qualitative sentiment data -words, gestures, and voice intonation of the traders and bankers – on top of collecting terra bytes of quantitative high frequency trading data- and then trying to unravel and learn how the two data sets will be fused to make a decision. Data fusion will be the next challenge, after the statistical machine learning folks have cleared the deep learning stable. I will talk about my experience of sentiment analysis and data fusion and will sprinkle my talk magical dust of artificial intelligence…more details

Talk | Deep Learning | Machine Learning | Intermediate

In this talk Laura will explain how to create an online image classifier by deploying deep learning models with the use of Flask and Docker. She will begin by giving an overview of some deep learning architectures and how models can be trained using the Keras library in order to extract a wealth of different information about the content of images. She will cover how to make inference calls to the trained models in order to obtain a probability and class label. There will be a walkthrough of some coded examples.

The talk will conclude by reflecting upon the advantages and disadvantages of using Docker in order to serve machine learning models versus a commercial production system…more details

Talk | Deep Learning | All Levels

This talk reviews a set of relevant deep learning model architectures including autoencoders, variational auto-encoders, generative adversarial networks and sequence-to-sequence methods, and addresses how they can be applied to the task of anomaly detection, comparing them in terms of training, inference and storage costs. Anomaly detection using each of these models is explored as a function of how they can be applied to first model normal behavior, and then this knowledge is exploited to identify deviations. In addition, we provide practical guidance for the successful implementation of anomaly detection systems within enterprises across key metrics like interpretability, reduction of false positives and scalability. We show benchmark results from applying them using several datasets and discuss industry best practices for deploying each model. Finally, we provide intuition for how and why these algorithms work by demoing them with a working prototype…more details

Talk | Deep Learning | All Levels

In this talk Dr Waseem Akhtar outlines ideas that employ the concept of Level of Abstraction (LoA) while finding meaning in information, to limit analysis of datasets to ensure privacy. LoA can be used to restrict people, organisations and algorithms so that while analysing datasets, information only at the permitted LoA is extracted; datasets are not analysed and mined for the information that is beyond the permitted LoA. It has widely been reported in literature that datasets are regularly mined to extract information that is well beyond the original, transparently stated purpose of capturing data. This may include finding patterns and personality traits using digital footprints of unsuspecting and unaware individuals. Locking a dataset to a particular, clearly described, thoroughly understood and mutely agreed LoA should ensure that the datasets are not used to extract information beyond clearly stated purpose for which data is originally captured…more details

Talk | Machine Learning | NLP | Intermediate-Advanced

The session will include a brief introduction on the concepts of Rare Events, Count Processes, Poisson and Negative Binomial Distribution Origins and Dynamic Time Warping. Following this, we will be discussing one area of application for this paper in great depth – starting from the problem formulation up to the final business impact + caveats. The entire algorithm will then be explained in the light of the problem statement (generalized alternates will also be given at every point). Since the algorithm is math heavy, we will be taking small detours to explain certain key concepts (such as Concords & Discords in time series, Relative Local Density Clustering, Outlierness, Correlation Mutation, Sister Classes, etc.) used in the solution; the audience will not be expected to know advance concepts before-hand – basic idea of time series forecast and distance measures would suffice. Finally, once the solution is laid down in a manner that can be easily replicated for other business solutions – we will discuss limitations and future work + (if time permits, a few more example business problems where the solution can be applied). This is a very generic algorithm applicable to cases in multiple disciplines: Retail, IT, Finance – basically anywhere a Rare Event is feasible in the form of a Time Series!..more details

Talk | Machine Learning | Beginner

The need for representing knowledge is ubiquitous in applications: travel apps need to know about transport links and accommodation; music apps need to know about artists and albums; (social) networking apps need to know about user profiles and preferences; and “virtual assistants” need to answer questions such as “what is the height of the Eiffel Tower”. Although real-world knowledge can be complex, a great deal can be achieved even with simple representations such as knowledge graphs; Google’s knowledge graph, for example, contains more than 50 billion facts, and enables Google to provide direct answers to many questions.

In this talk I will introduce basic knowledge graphs, illustrate the benefits of augmenting them with ontologies, explore some inherent challenges, and compare various techniques for query answering over ontology augmented large-scale knowledge graphs…more details

Talk | MLOps & Data Engineering | Stream Processing for Data Analytics | Beginner

Stream processing has fundamentally changed the way we build and think about data pipelines — but the technologies that unlock the value of this powerful paradigm haven’t always been friendly to non-Java/Scala developers.
Apache Flink has recently introduced PyFlink, allowing developers to tap into streaming data in real-time with the flexibility of Python and its wide ecosystem for data analytics and Machine Learning. In this talk, we will explore the basics of PyFlink and showcase how developers can make use of a simple tool like interactive notebooks to harness the full power of an advanced stream processor like Apache Flink...more details

Talk | Deep Learning | Machine Learning | Intermediate-Advanced

As the world rapidly changing around us, business across industries are treading carefully with unprecedented challenges and, if lucky enough, new opportunities. Due to their statistical assumption of generalizable patterns from the past, machine learning models are facing more scepticism about their validity in the world we now live in. It is more crucial than ever for data scientists to keep close eye on our beloved models in production, understand the impact of business changes on them, and steer promptly from potential pitfalls.

In this session, I will share some experience of model monitoring and diagnosis from a leading UK fintech company. We will discuss how to detect distributional change and analyse its impact on the model performance metrics. We will also look at how to decompose exogenous effect from vintage effect, which would help business understand the model validity in the long run. Finally, we will share some techniques to detect change of statistical relationship and discover new features.…more details

Talk | Machine Learning | Intermediate

Fuel usage and costs related to it are of enormous interest, especially in aeroespacial domain. In the past engineers had to manually go through thousands of flights in order to seek the abnormal ones. Positive anomalies are to be mimicked, for negatives it’s crucial to understand the reasons to avoid such scenarios in the future, normal flights help drawing insights about typical flight behavior. However, the process is tedious, time consuming and error prone as well as it’s hardly possible for a human to eyeball so large data volume. The flight anomaly detection project we delivered for our client made it possible to limit the manual work to a minimum while increasing effectiveness. Currently engineers only have to go through a limited subset of flights, just the suspicious ones proposed by the algorithm, so they have more time to create valuable insights and focus on other tasks...more details

Talk | Deep Learning | Machine Learning | Beginner-Intermediate

We’ve all heard that “a picture is worth a thousand words”. That definitely applies to our hotel price comparison service hotel.idealo.de, a daughter company of Axel Springer. Each hotel supplies us with dozens of images, presenting us with the challenge of choosing the most “attractive” image for each pitch on our offer comparison pages since photos can be just as important for bookings as reviews. Given millions of hotel offers, we end up with more than 100 million images which require an “attractiveness” assessment.

Although teaching and training a computer to decide what makes a “beautiful” hotel photo is a hard problem, it’s not impossible. In this talk, we will present how we solved this difficult problem. In particular, we will share our training approaches and the peculiarities of the models. We will also show the “little tricks” that were key to solving this problem…more details

Talk | Machine Learning | Beginner-Intermediate

Many machine learning models are opaque in the way they make a prediction. Even with quite common ensemble models such as random forest and gradient boosting, it is difficult to explain why the model made a particular decision. For certain business contexts, this poses a challenge. We faced such a hurdle when working on various machine learning models in the real estate domain. Among the things we were interested in predicting were the rent and price of a property, and duration of time during which it will be without a tenant if it were to become vacant. The business owners who were the users of the models were not satisfied with the common feature importance plot obtained with tree-based models.
In the talk I plan to briefly go over the business problem itself and the approach we took to solve it, as well as explain what Shapley value is, and how it can be used in many applications…more details

Talk | Deep Learning | Machine Learning | Intermediate

Despite technological advances in matching people with jobs, labor market is still highly inefficient: millions of jobs stay unfilled, while a comparable number of professionals are trapped in jobs they do not enjoy. Finding the right development path for a person or an organization requires an understanding of the current and future demands in terms of skills needed for every job. Current trends such as automation and robotization bring an added impact to anticipating talent development.

This talk will raise awareness on the importance of skills (hard and soft) in career progression. The audience will learn how career paths can be built using a skill-based approach. We will also share quantitative results on the impact of automation on skills, jobs, and job families. Finally, we will present a methodology and results on how organizations can estimate the training needs to upskill and reskill their workforce within a 3-5 year horizon...more details

Talk | Machine Learning | AI for Good | Intermediate-Advanced

Satellite imagery and machine learning can be used to estimate emissions from coal power plants across the world, and locate solar panels in the UK. Plentiful data, cloud compute, and open source tools means even small teams can do powerful things – in this talk we’ll go through the steps to setup a machine learning pipeline to analyse overhead imagery and discuss the results and lessons learnt.

For energy analysis, it is necessary to first deeply understand the distribution of the underlying infrastructure. Only then can the steps to create a representative and useful dataset be taken. There are many different imagery providers and datasets – the strengths and weaknesses of different data sources are discussed along with techniques to pre-process them.

Computer vision CNNs and transfer learning mean training a model on geospatial raster data in a pipeline is reasonably straightforward. This talk will go through the necessary steps for a generic task, and the code will be available…more details

Talk | Deep Learning | Machine Learning | Beginner-Intermediate

Doing the right thing has never been more important for people working in the areas of Data Science, Machine Learning and Artificial Intelligence. Data Ethics has been growing in importance recently and with recent global events has renewed the focus on the topic. This presentation will examine some of the main concepts involved in Data Ethics and will use a number of different case studies to illustrate the different challenges being encountered and how these can be address considering various legal aspects…more details

Talk | Data Visualization | All Levels

I learned that most valuable, essential skills to succeed in our business are not necessarily programming and statistics. In the business world is the ability to communicate – tell data-based stories, to visualize and to promote them is crucial. As a data scientist you can only be as good as your communications skills are, since you need to persuade others to make decisions or help to build products based on your analyses. Finally, dare to start simple. When you introduce data science in the industry, you start on the brown field. Simple use cases and projects like metrics, dashboards, reports, historical analysis help you to understand the business model and to assess where is your contribution to success of the company. This is the key to data science success, not only in the multimetal but everywhere else as well…more details

Talk | Data Visualization | Research Frontiers | Intermediate

Data visualization techniques based on readily-available libraries like d3.js and plotly can provide a highly effective means for exploring this underlying structure and facilitating better understanding. In our lab, we have produced multiple proof-of-concept multi-dimensional visualizations of complex datasets that can address the issues mentioned above.
This talk will illustrate selected use-cases that describe the process from design to deployment. Additionally, the talk will highlight our efforts in exploring different data visualization tools and engineering the best tool to help our diverse group of data managers, data curators, and researchers validate and select the right data for their analyses to produce more insightful papers that are enhanced by informative interactive visualizations…more details

Talk | Machine Learning | Computer Vision | Advanced

Clearly explaining a rationale for a classification decision to an end-user can be as important as the decision itself. Existing approaches for deep visual recognition are generally opaque and do not output any justification text; contemporary vision-language models can describe image content but fail to take into account class-discriminative image properties which justify visual predictions. In this talk, I will present my past and current work on Zero-Shot Learning, Vision and Language for Generative Modeling and Explainable Machine Learning where we show (1) how to generalize image classification models to cases when no visual training data is available, (2) how to generate images and image features using detailed visual descriptions, and (3) how our models focus on discriminating properties of the visible object, jointly predict a class label, explain why/not the predicted label is chosen for the image...more details

Talk | Machine Learning | NLP | Beginner-Intermediate

You will enjoy this talk if you’re interested in:

• Making the world more transparent
• Making Retail more sustainable
• Learning more about scraping and NLP

This talk will inform you about the general process and steps we have taken, as well as (technical) challenges and solutions, which should help you in setting up or improving your own open-source NLP project. Moreover, we will discuss possible improvements and encourage debate to further improve our project.

This year’s talk is aimed at raising awareness; if all goes well, we would like to organise a panel discussion at next year’s ODSC, to bring this topic even further...more details

Talk | Deep Learning | Machine Learning | Advanced

Machine Learning Systems play a huge role in several businesses from the Banking industry to recommender systems in entertainment applications until health domains. The era of “A Data Scientist with a Script in a single machine” is officially over in high stakes ML.

We’re entering an era of Machine Learning Operations (MLOps) where those critical applications that impact society and businesses need to be aware of aspects like active failures and latent conditions. This talk will discuss risk assessment in ML Systems from the perspective of reliability, safety, and especially causal aspects that can lead to the rise of silent risks in said systems…more details

Talk | Machine Learning | MLOps & Data Engineering | Intermediate

Model performance often takes precedence over model stability, if stability is considered at all, when optimising machine learning pipelines. While this can be beneficial in Kaggle competitions, it can lead to unexpected outcomes when a model is in production. For example, credit risk scores, medical classifiers, and customer-lifetime value models should be consistent over the same person. We present a real-world use case of how we optimise for both model stability and performance at Farfetch.

Farfetch is an online luxury-fashion platform with over one million active customers and more than one billion dollars of transactions in its marketplace yearly. We use machine learning to optimise customer relationship management (CRM) activities through customer-lifetime value and churn modelling. For this application, model stability is crucial for adoption by our internal stakeholders and to ensure a consistent customer experience…more details

Talk | AI in Business | Intermediate

Why “Small Data” is important for advancing the AI? Will Big Data models work correctly, if trained on synthetic data? Can we properly define and train a Neural Network from a limited dataset, by augmenting its data with an extra knowledge (meta-data)?
This keynote is about the role of small data in the future of AI. Efforts have already begun in this direction. Although the current mantra of deep learning says ‘you need big data for AI’, more often than not, AI becomes even more intelligent and powerful if it has the capability to be trained with small data. Some AI solutions that rely only on small data outperform those working with big data. Some other AI solutions use big data to learn how to take advantage of small data.
Presenter will illustrate on practical and visual examples from Fashion Retail sector, some approaches and methods for using Small and Synthetic Data for training of the AI and Big Data systems, starting from the “big difference” but common nature of Small Data and Big Data. It will be explained in details the case of a Visual Search service, who’s training is initially based on 3D CAD models and their meta data, combined with real pictures on the later stages of training…more details

Talk | Machine Learning | NLP | Beginner-Intermediate

On a daily basis, the Research, Selection & Communication team at Van Lanschot Kempen monitor the vast and ever-changing seas of financial news, concentrating its immense volume of information into a handful of well-researched articles written for the benefit of the bank’s investment clients. These clients receive such articles at the discretion of their advisor who – in an ideal world at least – is very familiar with their clients’ investment interests (e.g. tech stocks, companies with high ESG ratings, emerging markets, …), and therefore perfectly placed to recommend relevant news to them. However, with a growing number of clients to support, it becomes impossible for advisors to recall the preferences of each and every client. How then, can we ensure that Van Lanschot Kempen’s news content reaches those clients who’ll be most interested?

The key breakthroughs from this project which I’ll cover in this talk are:

– Leveraging in-house email and meeting note data to build a client profile
– Applying cutting-edge BERT pre-trained model to identify named entities relevant to the financial sector
– Productionalizing a recommender system to aide and guide traditional financial advisors...more details

Talk | Deep Learning | Machine Learning | Intermediate

We propose an end-to-end scalable solution using deep learning architecture consisting of a computer vision component connected to a sequence generation component. Through training on millions of documents, the model learns to understand document trends and characteristics to finally extract important fields from raw documents. There is marked improvement of accuracy compared to third-party OCR services. Additional benefits include character-level probabilities for confidence scores and using explainability algorithms such as LIME to determine which “hot pixels” in the document are responsible for the predictions…more details

Talk | Machine Learning | Quant Finance | Beginner-Intermediate

This talk will cover the following topics:

• The broad application of machine learning in finance

• Extracting sentiment from textual data such as news stories and social media content using machine learning algorithms

• Construction of scoring models and factors from complex data sets such as supply chain graph, options prices and ESG (Environmental, Social and Governance)

• Use of Alternative data such as extreme weather (Cyclone, Snowfall) to quantify the impact on companies that own retail stores and factories (geolocational).

• Robust portfolio construction by blending in factors derived from alternative data with traditional factors such as those in the Fama-French five-factor model.

• Machine Learning techniques for asset pricing, e.g. learning the complex quant models (PDE, Monte Carlo) via machine learning approach for efficient pricing of derivative securities…more details

Talk | Deep Learning | NLP | Intermediate

Forecasting the economy with fifty shades of emotions
This talk focuses on how news, emotions