• 08:00

    REGISTRATION & LIGHT BREAKFAST

  • 09:00
    Stephen O'Farrell

    WELCOME NOTE

    Stephen O'Farrell - Machine Learning Scientist - Bumble

    Down arrow blue

    BuzzWords - How Bumble does Multilingual Topic Modelling at Scale

    With the abundance of free-form text data available nowadays, topic modelling has become a fundamental tool for understanding the key issues being discussed online. We found the state-of-the-art topic modelling libraries either too naive or too slow for the amount of data a company like Bumble deals with, so we decided to develop our own solution. BuzzWords runs entirely on GPU using BERT-based models - meaning it can perform topic modelling on multilingual datasets of millions of data points, giving us significantly faster training times when compared to other prominent topic modelling libraries

    Stephen O’ Farrell is a machine learning scientist at Bumble, where, as a member of the Integrity & Safety team, he works to ensure user safety across all of Bumble’s platforms. His work generally deals with NLP and Computer Vision tasks - deploying deep learning models at scale across the organisation. He graduated with an MSc in Data Science and BSc in Computational Thinking, both from Maynooth University, Ireland

    Linkedin
  • IMPROVING DEEP LEARNING

  • 09:15
    Michael Bronstein

    Physics-Inspired Models for Deep Learning on Graphs

    Michael Bronstein - Head of Graph Learning Research / DeepMind Professor of AI - Twitter / University of Oxford

    Down arrow blue

    Physics-Inspired Models for Deep Learning on Graphs

    The message-passing paradigm has been the “battle horse” of deep learning on graphs for several years, making graph neural networks a big success in a wide range of applications, from particle physics to protein design. From a theoretical viewpoint, it established the link to the Weisfeiler-Lehman hierarchy, allowing to analyse the expressive power of GNNs. I argue that the “node and edge-centric” mindset of current graph deep learning schemes imposes insurmountable limitations that obstruct future progress in the field. As an alternative, I propose physics-inspired “continuous” learning models that open up a new trove of tools from the fields of differential geometry, algebraic topology, and differential equations so far largely unexplored in graph ML.

    Michael Bronstein is the DeepMind Professor of AI at the University of Oxford and Head of Graph Learning Research at Twitter. His research interests are primarily in geometric deep learning and graph ML. His work in these fields appeared in the international press and was recognised by multiple awards. Michael is a Member of the Academia Europaea, Fellow of IEEE, IAPR, BCS, and ELLIS, ACM Distinguished Speaker, and World Economic Forum Young Scientist. In addition to his academic career, he is a serial entrepreneur and founder of multiple startup companies, including Novafora, Invision (acquired by Intel in 2012), Videocites, and Fabula AI (acquired by Twitter in 2019).

    Twitter Linkedin
  • 09:40
    Edward Johns

    Humans Teaching Robots: The Future of Deep Learning in the Physical World

    Edward Johns - Director of the Robot Learning Lab / Head of Robot Learning - Imperial College London / Dyson

    Down arrow blue

    Humans Teaching Robots: The Future of Deep Learning in the Physical World

    Deep learning has proven to be astonishingly powerful for software-based AI. But what about the physical world? Could we use deep learning to train robots? In this talk, I will present my vision for the future of everyday robots learning everyday tasks, through interactions with everyday humans. In particular, I will describe the progress being made in the field of imitation learning – robots learning from human demonstrations – in the Robot Learning Lab at Imperial College London. I believe that robotics will perhaps change our lives more dramatically than any other area of AI, and that deep learning will be a key part of that future.

    Dr Edward Johns is the Director of the Robot Learning Lab at Imperial College London. After receiving a BA and MEng from Cambridge University, and a PhD from Imperial College, he was a founding member of the Dyson Robotics Lab at Imperial College in 2016. In 2017, he was awarded a prestigious Royal Academy of Engineering Research Fellowship, and then in 2018 he founded the Robot Learning Lab. In a part-time capacity, he is also Head of Robot Learning at Dyson, and an advisor for a number of start-ups. Edward's research lies at the intersection of robotics, computer vision, and machine learning.

    Twitter Linkedin
  • 10:05
    Ira Ktena

    Graph Representation Learning in Healthcare and Beyond

    Ira Ktena - Senior Researcher - DeepMind

    Down arrow blue

    Graph Representation Learning in Health Applications and Fairness Considerations

    Recent work on neuroimaging has demonstrated significant benefits of using population graphs to capture non-imaging information in the prediction of neurodegenerative and neurodevelopmental disorders. This has been enabled by advances in the field of graph representation learning. The non-imaging attributes may contain demographic information about the individuals, but also the acquisition site, as imaging protocols and hardware might significantly differ across sites in large-scale studies. This talk will give an overview of the advances that graph representation learning has contributed to the fields of neuroimaging and connectomics in recent years. It will also discuss fairness considerations that arise when these models leverage sensitive attributes.

    Ira is a Senior Researcher at DeepMind working on Machine Learning research for Life Sciences with Danielle Belgrave and the Deep Learning team. Previously, she was a senior Machine Learning Researcher with the Cortex Applied Research team at Twitter UK, focusing on real-time personalisation while she carried out research at the intersection of recommender systems and algorithmic transparency. Her exploration on algorithmic amplification of political content on Twitter was featured by the Economist and the BBC, among others.

    Twitter Linkedin
  • 10:30

    COFFEE & NETWORKING BREAK

  • TOOLS FOR DEEP LEARNING

  • 11:15
    Tom Sadler

    Delivering The Right Technologies at the Right Time

    Tom Sadler - Data Science, AI &Edge Solution Lead for UK&I, Advanced Compute & Solutions - HP Inc

    Down arrow blue

    Delivering The Right Technologies at the Right Time

    Machine learning helps organisations extract value from data but not without cost. Creating an efficient path to value for AI requires agility in where, and how, enterprise data science and machine learning teams work. Machine learning models are the heart of AI and to make more accurate predictions they must be trained and optimised over many iterations, typically with vast amounts of data. They are often conceptual as teams tweak their approach to solve a problem. Increased iterations lead to higher model quality, which directly improves both revenue and margins for companies. Hear from HP about the Z by HP high performance workstations together with NVIDIA Quadro GPUs that meet the demands of advanced analytics and enable customers like American Airlines to fuel innovations.

    Tom joined HP earlier in the year to manage HP Inc’s Data Science and AI business for the UK&I. He works with some of the UK&I’s biggest organisations across Private and Public sectors, helping them implement HP’s Data Science and Edge Solutions.Before working for HP Tom has worked in the IT industry for the last decade working with some of the biggest organisations in the UK and Europe. Providing and applying Technology solutions.

    Linkedin
  • 11:45

    PANEL: Harnessing the Power of Deep Learning

  • Chanuki Seresinhe

    MODERATOR

    Chanuki Seresinhe - Head of Data Science - Zoopla

    Down arrow blue

    Dr. Chanuki Illushka Seresinhe is currently the Head of Data Science at Zoopla, managing data scientists at both Zoopla and Hometrack. Chanuki began working as a commercial data scientist at Channel 4 and, more recently, was the Director of Data Science for Culture Trip.

    Chanuki has a PhD in Data Science from the University of Warwick. Her research at the University of Warwick and the Alan Turing Institute involves using big online datasets and deep learning to understand how the aesthetics of the environment influences human wellbeing. Her research has been featured in the press worldwide including the Economist, Wired, The Times, BBC, Spiegel Online, Guardian, Telegraph, Scientific American, Newsweek and MIT Technology Review.

    Twitter Linkedin
  • Hongyu Li

    PANELIST

    Hongyu Li - Senior Data Scientist - Virgin Media o2

    Down arrow blue

    Dr. Hongyu Li received his Ph.D in Machine Learning from the University of Sheffield in 2007. After that, he spent 8 years in Ericsson as a network consultant in various areas including network modelling. He then joined Hutchison 3G (Three) as a network strategy specialist in 5G transport network development. He joined Telefonica UK (O2, now Virgin Media O2) in 2019 to lead advanced data analytics programs.

    Linkedin
  • Abdullahi Adamu

    PANELIST

    Abdullahi Adamu - Senior Software Engineer - Sony

    Down arrow blue

    Abdullahi Adamu got his PhD from the University of Nottingham in Computer Science (specialising in neuroevolution of efficient Artificial Neural Networks using neural diversity) in 2016. Since then he has worked in various industries including market research and advertising. These included working at a market research startup in London where he developed models that could extract insights from human conversations about products or services. He moved on to Publicis, where he wore the hat of a Data Engineer and Data Scientist in 2018, and contributed by writing a framework to aggregate campaign data from various campaign managers that allows for seamlessly querying through a dashboard. He also worked on using deep learning projects to various problems; these included questions like why certain ads performed better than others for different audiences? how visible are sponsorship ads? and how are inclusive video ads in terms of age, gender, ethnicity and disability.

    Outside of work, apart from Netflix; he loves watching movies in the cinema, especially if it's directed by Tarantino, Christopher Nolan or Peter Jackson. He enjoys playing around with IoT devices to make everyday items “smart” and loves traveling to discover new places, cultures and make new friends.

    Linkedin
  • 12:35

    LUNCH

  • 13:50
    Zhenwen Dai

    Reinforcement Learning at Spotify: An Example with Interactive Radio

    Zhenwen Dai - Staff Research Scientist - Spotify

    Down arrow blue

    Reinforcement Learning at Spotify: An Example with Interactive Radio

    At Spotify, we are developing reinforcement learning based personalization experience. In this talk, I will present an example of our reinforcement learning development: a Radio playlist experience powered by reinforcement learning, which is called Interactive Radio. In an Interactive Radio listening session, a reinforcement learning agent refreshes the radio station upon user feedback. This is different from the typical ranking paradigms in recommender system literature. In this talk, I will show the simulation-based approach that we developed to tackle this problem and discuss the results of our approach from online tests.

    Zhenwen Dai is a staff research scientist at Spotify. Zhenwen’s research interest is to develop machine learning systems that can automatically learn from large amounts of unlabeled data and make informed decisions. Prior to this, he was a machine learning scientist at Amazon in Cambridge, UK. He did his PhD in machine learning at Goethe University Frankfurt and was followed by a postdoc at Sheffield University.

    Linkedin
  • 14:15
    Detlef Nauck

    How to Play Fair. Fairness Assurance for AI Models

    Detlef Nauck - Head of AI & Data Science Research - BT

    Down arrow blue

    Implementing a Company-Wide Framework for Responsible AI

    The BT Group Manifesto explains how our responsible tech principles aim to ensure that while our tech is commercially viable and profitable, it is always for good, accountable, fair and open. When we build AI solutions we follow a company-wide governance framework that helps us to build AI ethically and supports developers in following best practice. The framework is underpinned by BT’s AI research programme and I will present results from our work on fairness evaluation of machine learning models and model monitoring.

    Detlef Nauck is the Head of AI & Data Science Research for BT’s Applied Research Division located at Adastral Park, Ipswich, UK. Detlef has over 30 years of experience in AI and Machine Learning and leads a programme spanning the work of a large team of international researchers who develop capabilities underpinning modern AI systems. A key part of Detlef’s work is to establish best practices in Data Science and Machine Learning for conducting data analytics professionally and responsibly. Detlef has a keen interest in AI Ethics and Explainable AI to tackle bias and to increase transparency and accountability in AI. Detlef is a computer scientist by training and holds a PhD and a Postdoctoral Degree (Habilitation) in Machine Learning and Data Analytics. He is a Visiting Professor at Bournemouth University and has published 3 books, over 120 papers, and holds over 20 AI patents.

    Linkedin
  • COMPUTER VISION

  • 14:40
    Jerome Pasquero

    Frequently Asked Data Labeling Questions

    Jerome Pasquero - Senior Product Manager - Sama

    Down arrow blue

    Frequently Asked Data Labeling Questions

    A day in the life of an ML engineer or a data scientist is not as glamorous as you might think; data-related tasks — from aggregating to labeling and augmenting data — can take up to 80% of their time. At Sama, we’ve helped hundreds of organisations overcome data challenges at every stage of the AI model lifecycle. In this session, we will cover the most frequently asked questions about data labeling. Key topics include: - Data curation - Representation and bias in data - Edge cases and ambiguity - Performance of models in production

    As Sr Product Manager at Sama, Jerome Pasquero leads the expansion of Sama's AI Product & Solutions portfolio. Jerome brings extensive experience leading multiple AI product & technology initiatives to successfully introduce innovative AI ML solutions in the market. Jerome has developed leading-edge technologies that range from pure software applications to electromechanical devices while being a key contributor to the design of innovative and successful consumer products that have shipped to millions of users. Jerome is also listed as inventor on more than 120 US patents and has published over 10 peer-reviewed journal and conference articles. Jerome holds a Ph.D. in electrical engineering from McGill University in Montreal, Canada.

    Twitter Linkedin
  • 15:05

    COFFEE & NETWORKING BREAK

  • 15:45
    Xin Wang

    Machine Vision at Shell

    Xin Wang - Machine Vision Manager - Shell

    Down arrow blue

    Machine Vision at Shell

    Shell is moving rapidly on digital transformation, in which AI plays a key role. In Shell we are developing and delivering a lot of computer vision projects/products. For this presentation, we will demo the latest development of Computer Vision in Shell and highlight some Computer Vision projects. All these developments are bringing values to Shell.

    Xin Wang was graduated from Delft University of Technology in 2015 with PhD thesis titled “Active Vision for Humanoid Robots”. Afterwards, she joined Shell and established Machine Vision team. Now she is Machine Vision manager and leads a team to deliver Machine Vision products to business. She has great passion in AI not only limited to Machine Vision, but also in the areas of Machine Learning, Natural Language Processing. In her spare time, she is active in teaching robotics to kids.

    Linkedin
  • Ahmed Hussein

    Machine Vision at Shell

    Ahmed Hussein - Senior Machine Vision Specialist - Shell

    Down arrow blue

    Machine Vision at Shell

    Shell is moving rapidly on digital transformation, in which AI plays a key role. In Shell we are developing and delivering a lot of computer vision projects/products. For this presentation, we will demo the latest development of Computer Vision in Shell and highlight some Computer Vision projects. All these developments are bringing values to Shell.

    Ahmed has been in shell for 3.5 years, looking after a portfolio of machine vision applications, from ideation, proofing concepts to piloting and deployment. He also works on strategic enablers and infrastructure building to accelerate the delivery of MV solutions at Shell. Before joining Shell, he had a mixture of academic and industry experiences, focusing mainly on deep learning and intelligent agents.

    Linkedin
  • 16:10
    Huma Lodhi

    Deep Learning Techniques to Improve the Football Viewing Experience

    Huma Lodhi - Lead Machine Learning Engineer - Sky

    Down arrow blue

    Computer Vision Meets Deep Learning: A Sports Fan Perspective

    Sports played an important role in the development of Artificial Intelligence and Machine Learning. Chess has been an attractive application domain since the early days of AI due to intelligence and reasoning required to play this. Recently there has been an interest in applying AI and more specifically deep learning techniques for generating solutions for tasks ranging from interesting event detection to predicting results to enhance viewer’s experience and increase their engagement. This talk will give an overview of novel methodologies based on deep learning and computer vision for sports from a viewer’s perspective.

    Huma Lodhi is the Lead Machine Learning Engineer at Sky. She has over 15 years of experience in Artificial Intelligence & Machine Learning across both industry and academia. She is an accomplished expert with hands on experience in development and application of Deep Learning, Kernel Methods, Relational Learning and Ensemble Methods for areas ranging from insurance to health care. She has a PhD in Machine Learning from university of London. She is a co-editor of two books and has published many research articles in leading AI & Machine Learning journals and conferences.

    Linkedin
  • REINFORCEMENT LEARNING

  • 16:35
    Vincent Moens

    TorchRL: The PyTorch Reinforcement Learning Domain Library

    Vincent Moens - Applied Machine Learning Research Scientist - Meta

    Down arrow blue

    TorchRL: The PyTorch Reinforcement Learning Domain Library

    We present TorchRL, the new reinforcement learning library from the PyTorch ecosystem team. TorchRL is an open-source Reinforcement Learning (RL) library for PyTorch. It provides pytorch and python-first, low and high level abstractions for RL that are intended to be efficient, documented and properly tested. The code is aimed at supporting research in RL. Most of it is written in python in a highly modular way, such that researchers can easily swap components, transform them or write new ones with little effort. Given reinforcement learning is a nascent space and the largest production applications are likely to come from tomorrow’s research, we deliberately focus on supporting emerging areas. Through simple examples, we illustrate the capabilities of this library and its ease of use in various RL subfields.

    Vincent Moens is a research engineer at Meta working on developing the TorchRL library — the Reinforcement Learning ecosystem library for PyTorch. TorchRL will be an open-source library built on top of PyTorch to support research in RL through a set of low and high-level reusable primitives that are common across RL frameworks. In 2013 Vincent graduated from Med School in Brussels and during the course of his residency in neurology he undertook a PhD in cognitive and computational neuroscience at UCLouvain, Belgium.

    After completing his PhD, he worked in the financial sector as a Machine Learning Scientist for a couple of years in London. He then switched back to research and worked as a Senior Machine Learning Research Scientist at Huawei, where he integrated the reinforcement learning team and provided expertise in generative modeling for model-based reinforcement learning.

    Twitter Linkedin
  • 17:00

    NETWORKING RECEPTION

  • 18:00

    END OF DAY 1

  • 08:00

    DOORS OPEN & LIGHT BREAKFAST

  • 09:00
    Stephen O'Farrell

    WELCOME NOTE

    Stephen O'Farrell - Machine Learning Scientist - Bumble

    Down arrow blue

    BuzzWords - How Bumble does Multilingual Topic Modelling at Scale

    With the abundance of free-form text data available nowadays, topic modelling has become a fundamental tool for understanding the key issues being discussed online. We found the state-of-the-art topic modelling libraries either too naive or too slow for the amount of data a company like Bumble deals with, so we decided to develop our own solution. BuzzWords runs entirely on GPU using BERT-based models - meaning it can perform topic modelling on multilingual datasets of millions of data points, giving us significantly faster training times when compared to other prominent topic modelling libraries

    Stephen O’ Farrell is a machine learning scientist at Bumble, where, as a member of the Integrity & Safety team, he works to ensure user safety across all of Bumble’s platforms. His work generally deals with NLP and Computer Vision tasks - deploying deep learning models at scale across the organisation. He graduated with an MSc in Data Science and BSc in Computational Thinking, both from Maynooth University, Ireland

    Linkedin
  • UNSUPERVISED/SEMI-SUPERVISED LEARNING

  • 09:10
    Simone Foti

    Latent Disentanglement for the Generation of 3D Digital Humans and Plastic Surgery Applications

    Simone Foti - PhD in Geometric Deep Learning - UCL

    Down arrow blue

    Latent Disentanglement for the Generation of 3D Digital Humans and Plastic Surgery Applications

    The generation of 3D human faces and bodies is a complex task with multiple potential applications ranging from movie and game productions, to augmented and virtual reality, as well as metaverse applications. However, learning a disentangled, interpretable, and structured latent representation in 3D generative models is still an open problem and state-of-the-art methods for latent disentanglement are not able to disentangle identity attributes of faces and bodies. This talk will give an overview of recent self-supervised approaches to train a 3D shape variational autoencoder and encourage a disentangled latent representation of identity attributes. In addition, it will discuss how these methods can improve the diagnosis of craniofacial syndromes and aid surgical planning.

    Simone is finishing a PhD at the University College London (UCL). His research lies at the intersection of geometric deep learning, computer vision, and computer graphics and aims at developing new latent disentanglement techniques to improve character generation and shape analysis. During the PhD, he did internships at Disney Research Studios and Adobe Reserch, where he researched single-image 3D face reconstruction methods and implicit functional representations for patch-driven super-resolution of textures.

    Linkedin
  • 09:35
    Krisztina Sinkovics

    Semi-Supervised Object Detection for Agricultural Robotics

    Krisztina Sinkovics - AI Research Engineer - Small Robot Company

    Down arrow blue

    Semi-Supervised Object Detection for Agricultural Robotics

    Producing more with less is a key challenge for the future of food production. Small Robot Company is reimagining farming to bring sustainability and profitability together. Our autonomous survey robots generate massive amounts of data from the fields. To make the most of this data we develop semi-supervised Deep Learning models that go beyond the typical focus on COCO and Pascal datasets. Our AI team is dedicated to sharing our research with the wider ML community. To this end we have released a public dataset on crop establishment and open-sourced a framework for semi-supervised object detection suited for any dataset.

    Krisztina Sinkovics is AI Research Engineer at Small Robot Company. She has been working on ML and DL research and engineering for the past six years, tackling applications in a multitude of areas ranging from ultrasonic non-destructive testing and procurement optimization to per-plant farming and application of GANs in geophysics. She holds a degree in economics with a focus on statistical modelling. Her recent work combines robotics, computer vision and representation learning under semi-supervised and unsupervised settings

    Linkedin
  • RESPONSIBLE AI

  • 10:00
    Toju Duke

    Responsible AI at Google

    Toju Duke - Program Manager - Responsible AI - Google

    Down arrow blue

    Responsible AI at Google

    AI is fundamental, groundbreaking technology with an adoption rate of 64% year over year. Along with its innovative and transformational abilities comes the challenges it faces with regards to ethics and responsibility. If AI/ML systems are developed without responsible and ethical frameworks, they have the propensity to deploy harm amongst individuals within society. It is the responsibility of every organisation developing AI models, to have a Responsible framework they adhere to which is accountable, fair, transparent and safe. In this talk, you’d learn how Google approaches Responsible AI, best practices for AI frameworks and relevant case studies.

    Toju is a Responsible AI Program Manager at Google, with over 15 years experience spanning across Advertising, Retail, Not-For Profits and Tech. She designs Responsible AI programs focused on the development and implementation of Responsible AI frameworks amongst Google’s product areas, with a focus on Foundation Models, Natural Language Processing, and Generative Language Models. With a proven track record on business success and project management, she is a Manager for Women in AI Ireland, Tech start-ups Mentor, and a Business Advisor. Toju is a public speaker and advocates for transparent and bias free AI aimed at reducing systemic injustices and furthering equality. She is also the founder of VIBE, a women's community focused on personal and professional development using the underlying principles of emotional intelligence.

    Twitter Linkedin
  • 10:25

    COFFEE & NETWORKING BREAK

  • 10:55
    Emmanuel Ferreyra Olivares

    Reliable AI Models: How to Deal With the Unknown?

    Emmanuel Ferreyra Olivares - Principal Researcher - Fujitsu Research of Europe

    Down arrow blue

    Reliable AI Models: How to Deal With the Unknown?

    Improving the reliability and robustness of modern AI models have received close attention lately due to its importance for critical applications. A significant challenge relates to the lack of indications from unknown data. As a result, models can produce an overconfident prediction on unseen inputs leading to erroneous model outcomes. This situation impacts model reliability resulting in compromises concerning security, monetary, competitiveness and eventually trust. Out-of-Distribution (OOD) research has focused on mitigating this problem by enabling detection techniques covering different fronts of this issue. This talk introduces the OOD concept and its implications on models’ security. Subsequently, it highlights relevant use cases in which the impact of OOD can be observed. Finally, the presentation concludes by pointing out the latest advances and the shortcomings still to be solved in future research directions.

    Emmanuel Ferreyra Olivares, PhD, is an AI & Data Security researcher with Fujitsu Research of Europe. In this role, Emmanuel is involved in designing and developing reliable and secure data-driven AI solutions relevant for highly regulated industries by practising close collaboration with global partners both in the industry and academia. Emmanuel’s research interests are in the broad areas of Cyber Security, Explainable AI, Computational Intelligence and Smart Simulation.

    Linkedin
  • APPLIED DEEP LEARNING

  • 11:20
    Hooman Shayani

    How Deep Learning is Revolutionising Design and Creativity

    Hooman Shayani - Senior Principal AI Research Scientist and Research Manager - Autodesk

    Down arrow blue

    How Deep Learning is Revolutionising Design and Creativity

    The accelerating impact of Deep Learning on creative professions in arts, engineering, and sciences surprises even the ML experts these days. We have already witnessed the sudden emergence of text-to-image models changing the landscape of visual arts. Deep Learning is poised to revolutionise the way we imagine, design, and make everything. Given so many challenges we are facing on this planet, that is mostly a good thing. With different examples from recent research at Autodesk for different industries such as Design, Architecture, and Engineering, I demonstrate how Deep Learning can not only enable computers to take care of the repetitive and mundane tasks in many creative jobs, but also inspire and augment the creativity and imagination of the creators.

    Hooman Shayani is a Senior Principal AI Research Scientist and Research Manager at Autodesk AI Lab. He is also an Honorary Professor at UCL CS Department. He joined Autodesk Research when Autodesk acquired one of the start-ups he has co-founded. Since then, he focused his research on generative design, and application of deep learning in design and manufacturing. He has papers and patents on Creative AI, generative models, spiking neural networks, evolutionary and bio-inspired computing, evolvable hardware, additive manufacturing, and generative design.

    Twitter Linkedin
  • 11:45
    Stephen O'Farrell

    BuzzWords - How Bumble does Multilingual Topic Modelling at Scale

    Stephen O'Farrell - Machine Learning Scientist - Bumble

    Down arrow blue

    BuzzWords - How Bumble does Multilingual Topic Modelling at Scale

    With the abundance of free-form text data available nowadays, topic modelling has become a fundamental tool for understanding the key issues being discussed online. We found the state-of-the-art topic modelling libraries either too naive or too slow for the amount of data a company like Bumble deals with, so we decided to develop our own solution. BuzzWords runs entirely on GPU using BERT-based models - meaning it can perform topic modelling on multilingual datasets of millions of data points, giving us significantly faster training times when compared to other prominent topic modelling libraries

    Stephen O’ Farrell is a machine learning scientist at Bumble, where, as a member of the Integrity & Safety team, he works to ensure user safety across all of Bumble’s platforms. His work generally deals with NLP and Computer Vision tasks - deploying deep learning models at scale across the organisation. He graduated with an MSc in Data Science and BSc in Computational Thinking, both from Maynooth University, Ireland

    Linkedin
  • 12:10
    Alberto Rizzoli

    Errors in Training Data: How to Spot Them and Their Effect on Model Performance

    Alberto Rizzoli - Co-founder & CEO - V7

    Down arrow blue

    Errors in Training Data: How to Spot Them and Their Effect on Model Performance

    Misclassified objects, loose bounding boxes, overlapping mask classes - How badly do they affect your AI? Lyft’s Level 5 dataset was found to have missing objects in 70% of its data, some ImageNet classes are up to 92% wrong, and 8 out of 10 ML teams change their label schema between their first model and their production release. How do we stay on top of training data errors, and how do they affect AI deployments in enterprise? We’ll explore examples of how bad training data led to incorrect business results, how to spot errors in your datasets, and how to fix them. This talk will cater to both business and technical audiences, showcasing both qualitative and quantitative results of introducing “bad data” into various computer vision domains.

    Alberto Rizzoli is co-Founder and CEO of V7, a platform for deep learning teams to manage training data workflows and create image recognition AI. V7 is used by over 300 global AI companies and enterprises including GE, Fujifilm, Merck, and MIT.

    Alberto founded his first startup at age 19 becoming MakerFaire’s 20under20. In 2015 founded Aipoly with Simon Edwardsson the first engine capable of running large deep neural networks on smartphones, leading to the creation of an app enabling the blind identify 5,000 objects through their phone camera used over 3 billion times.

    Today he leads V7, one of the UK's fastest growing startups powering the computer vision of millions of healthcare devices, robots, and self-driving cars.

    Alberto's work on AI granted him an award and personal audience by Italian President Sergio Mattarella, as well as Italy’s Premio Gentile for Science and Innovation. V7's underlying technology won the CES Best of Innovation in 2017 and 2018.

    Linkedin
  • 12:25

    LUNCH

  • 13:25
    Niraj Kumar

    Quantum Machine Learning and Its Impact on Finance

    Niraj Kumar - Quantum Algorithm Research Lead - JP Morgan

    Down arrow blue

    Quantum Machine Learning and Its Impact on Finance

    Quantum computing is one of the most promising technological advancements of our current generation with a promise to revolutionize many industries including pharma, automobile, finance, and cryptography among others. A recent surge of research activities within this field is to build quantum machine learning algorithms with the ability to run on the near term small scale noisy quantum computers with the prime objective of achieving a quantum advantage i.e. achieving the level of solution that a classical machine learning model would not be able to achieve. In this talk, I will give an overview of the quantum machine learning approaches for generative modeling and supervised learning and see how they impact quantitative finance. Specifically we will talk about building generative quantum models and how to speed up Monte-Carlo sampling.

    Dr. Niraj Kumar is a Vice President, Quantum Algorithm Research Lead at JP Morgan Chase&Co. His interests and active research work span the fields of quantum machine learning, verification of quantum devices, and secure quantum communications. He obtained his Ph.D. from Telecom Paristech, Paris where he worked on developing secure quantum communication protocols. He has an avid research interest in quantum for finance and has previously worked with PayPal as the head of Quantum Algorithms, and with PASQAL where he developed quantum algorithms for fraud detection and market forecasting, among other applications.

    Linkedin
  • 13:50
    Hastagiri Vanchinathan

    Billion Scale Recommendations at Sharechat and Moj

    Hastagiri Vanchinathan - Senior Director of AI - ShareChat

    Down arrow blue

    Billion Scale Recommendations at Sharechat and Moj

    Content marketplaces (like Sharechat, Moj, Instagram, Tiktok) face unique challenges in recommending content to their users. In addition to traditional end user metrics, these recommender systems will also have to care heavily about fairness and equity of the content creators. The volume of content that is uploaded to these platforms far exceeds the traditional commerce, music or movies use cases by orders of magnitude. For instance, on Moj the number of uploads per hour by our content creators is approximately equal to the total number of content pieces (including individual episodes) on Netflix historically. Making matters harder on short video platforms such as Moj, the average length of a content piece is around 15-20 secs while the average session lasts more than 30 minutes. This means that traditional recommender systems that care purely about getting the top 5-10 recommendations absolutely correct are not going to work very well in this case as we need to continue to maintain relevance and interest well into the top 200-300 recommendations. In this talk, I will give a brief overview of the AI journey at Sharechat and Moj - the number 1 Indian content marketplace platform. I will present some of the research challenges we are solving along with techniques and results that worked for us. I will also talk about some of the key decisions that we took along the way that helped us scale up AI org and its efficiency. The talk will have a mix of technical, research and strategic discussion points in our journey so far.

    Hastagiri is the Senior Director of AI at ShareChat, India's largest AI powered content ecosystem driven largely by feed personalisation, automated content understanding and improvements in camera and creator tools!

    Linkedin
  • 14:15

    PANEL: From Research to Reality: What is Next for Deep Learning?

  • MODERATOR

  • Abdullahi Adamu

    MODERATOR

    Abdullahi Adamu - Senior Software Engineer - Sony

    Down arrow blue

    Abdullahi Adamu got his PhD from the University of Nottingham in Computer Science (specialising in neuroevolution of efficient Artificial Neural Networks using neural diversity) in 2016. Since then he has worked in various industries including market research and advertising. These included working at a market research startup in London where he developed models that could extract insights from human conversations about products or services. He moved on to Publicis, where he wore the hat of a Data Engineer and Data Scientist in 2018, and contributed by writing a framework to aggregate campaign data from various campaign managers that allows for seamlessly querying through a dashboard. He also worked on using deep learning projects to various problems; these included questions like why certain ads performed better than others for different audiences? how visible are sponsorship ads? and how are inclusive video ads in terms of age, gender, ethnicity and disability.

    Outside of work, apart from Netflix; he loves watching movies in the cinema, especially if it's directed by Tarantino, Christopher Nolan or Peter Jackson. He enjoys playing around with IoT devices to make everyday items “smart” and loves traveling to discover new places, cultures and make new friends.

    Linkedin
  • Krisztina Sinkovics

    PANELIST

    Krisztina Sinkovics - AI Research Engineer - Small Robot Company

    Down arrow blue

    Semi-Supervised Object Detection for Agricultural Robotics

    Producing more with less is a key challenge for the future of food production. Small Robot Company is reimagining farming to bring sustainability and profitability together. Our autonomous survey robots generate massive amounts of data from the fields. To make the most of this data we develop semi-supervised Deep Learning models that go beyond the typical focus on COCO and Pascal datasets. Our AI team is dedicated to sharing our research with the wider ML community. To this end we have released a public dataset on crop establishment and open-sourced a framework for semi-supervised object detection suited for any dataset.

    Krisztina Sinkovics is AI Research Engineer at Small Robot Company. She has been working on ML and DL research and engineering for the past six years, tackling applications in a multitude of areas ranging from ultrasonic non-destructive testing and procurement optimization to per-plant farming and application of GANs in geophysics. She holds a degree in economics with a focus on statistical modelling. Her recent work combines robotics, computer vision and representation learning under semi-supervised and unsupervised settings

    Linkedin
  • Alberto Rizzoli

    PANELLIST

    Alberto Rizzoli - Co-founder & CEO - V7

    Down arrow blue

    Errors in Training Data: How to Spot Them and Their Effect on Model Performance

    Misclassified objects, loose bounding boxes, overlapping mask classes - How badly do they affect your AI? Lyft’s Level 5 dataset was found to have missing objects in 70% of its data, some ImageNet classes are up to 92% wrong, and 8 out of 10 ML teams change their label schema between their first model and their production release. How do we stay on top of training data errors, and how do they affect AI deployments in enterprise? We’ll explore examples of how bad training data led to incorrect business results, how to spot errors in your datasets, and how to fix them. This talk will cater to both business and technical audiences, showcasing both qualitative and quantitative results of introducing “bad data” into various computer vision domains.

    Alberto Rizzoli is co-Founder and CEO of V7, a platform for deep learning teams to manage training data workflows and create image recognition AI. V7 is used by over 300 global AI companies and enterprises including GE, Fujifilm, Merck, and MIT.

    Alberto founded his first startup at age 19 becoming MakerFaire’s 20under20. In 2015 founded Aipoly with Simon Edwardsson the first engine capable of running large deep neural networks on smartphones, leading to the creation of an app enabling the blind identify 5,000 objects through their phone camera used over 3 billion times.

    Today he leads V7, one of the UK's fastest growing startups powering the computer vision of millions of healthcare devices, robots, and self-driving cars.

    Alberto's work on AI granted him an award and personal audience by Italian President Sergio Mattarella, as well as Italy’s Premio Gentile for Science and Innovation. V7's underlying technology won the CES Best of Innovation in 2017 and 2018.

    Linkedin
  • 15:00

    CLOSING COFFEE BREAK

  • 15:30

    END OF SUMMIT

This website uses cookies to ensure you get the best experience. Learn more