Research Group, Computer Science, University of Hull
We are the Big Data Analytics research group (BDA) in the School of Computer Science at the University of Hull.
We collaborate with multiple other research groups in Computer Science, such as the Artificial Intelligence and Machine Learning group, the IoT group, and the Dependable Intelligent Systems (DEIS) group.
We also work closely with other parts of the university, including the
Viper HPC team, the
Energy and Environment Institute, the
Aura Centre for Doctoral Training Centre in Offshore Wind Energy and the Environment,
and the Centre of Excellence in Data Science, Artificial Intelligence and Modelling (DAIM).
Our academic partners involve the Universities of Sheffield, Lincoln, Durham, Strathclyde, Manchester and the Open University, amongst others.
Our non-academic partners involve businesses such as Reckitt, Toshiba, IBM, Spencer and Connexin, as well as
public sector organisations, such as the ORE Catapult, CEFAS and
the Environment Agency.
We first came together as a group in 2017 and have since been growing an interdisciplinary research portfolio in artificial intelligence and
data analytics, particularly focusing on research in (deep) machine learning, interactive systems and natural language processing, and increasingly
AI for sustainability and environmental modelling. We are keen to collaborate with partners in different areas to generate novel fundamental research, knowledge transfer and impact acceleration projects. This page is our home!
We apply machine learning, mostly deep learning, in all our research projects, due to its ability to learn features from large datasets and identify patterns in disparate, large and noisy datasets. Most of our work on natural language processing and other time-series applications uses LSTM/GRU or transformer models (e.g. for sentiment analysis, natural language generation) while other projects in image analysis and sensory output analysis have used CNNs or autoencoders.
Our research in this area focuses on deep learning models that are transparent and interpretable and can explain the rationales behind their actions - this is important in many commercial and real-life applications to increase trust, helps to avoid bias and discrimination in AI models, and makes them generally more accountable.
We are also interested in fast learning systems and investigate topics in transfer learning, divide-and-conquer approaches to learning and efficiency. See our deep learning publications for details.
Natural language processing (NLP) refers to approaches that analyse, generate and process natural language automatically using computational techniques and algorithms, mostly based on machine learning. Our strengths lie in natural language generation, that is the automatic generation of language from non-linguistic data sources, such as data-to-text generation, (spoken) dialogue systems, i.e. systems that engage in conversations with users to solve problems or provide information, sentiment analysis in text and social media and text classification and analysis.
We are also interested in social media analysis, including projects on sentiment analysis and mental health based on linguistic profiling, flood forecasting and emergency situation assessment from social media and topics in digital conservation, such as public opinion mining and the monitoring of species. See our natural language processing publications for details.
Interactive systems can interact with their users in different ways, including text, speech, haptic or other multimodal input. A common distinction is between task-oriented interactive systems that target a particular domain, e.g. answering user queries on restaurant or tourist attractions in the local area, and open-domain systems who attempt to answer any question, e.g. using information retrieval methods from the web. Most personal assistant fall into the latter category. Finally, chatbots have no purpose of informing but rather aim to entertain users by engaging in casual chat.
Our group is interested in human-like behaviour of interactive systems, e.g. the incremental planning of utterances and the analysis of linguistic features in different conversational contexts. Most recently we have worked on interactive decision support systems for various real-life applications, most notably in the wind domain, where we use interaction as a way to provide easy and intuitive access to the inner workings of wind turbines to support operations and maintenance. See our interactive systems publications for details.
We are applying AI and machine learning techniques to help solve problems in sustainability and environmental research. We are part of the EPSRC Aura CDT on Offshore Wind and the Environment, and we have a strong interest in the digitalisation of offshore wind. As part of this, we use AI to improve operations and maintenance of wind turbines to improve their reliability and usability as an alternative to fossil fuel-based energy sources. We achieve this through detailed and explainable fault forecasting and the automatic generation of maintenance strategies from expert knowledge sources.
Other research includes the forecasting of floods from high-frequency and social media data sources and the monitoring of water quality in the North Sea from sensory data and satellite images to achieve longer-term forecasting of the effects of human activity or environmental factors on water quality. We are currently growing our research on using AI towards the conservation of species. See our sustainability publications for details.
£5.8 EPSRC-funded CDT with £12M industry funding for at least 65 PGR students. Led by Hull with partners Durham, loughborough and Sheffield.
2024-32, Dethlefs (CI)
Project on Mortality Risk Prediction of ICU Patients with Sepsis Considering Dynamic Time Series Characteristics Under Uncertainty, with Chao Huang.
2023-25, Ma
We are extremely excited to have been awarded an Innovation Launchpad Network+ grant for Researchers in Residence. We will work with OREC to consolidate and expand fundamental research on using natural language to enhance O&M in OSW.
UKRI, 2023-25, Dethlefs (PI)
We are working with broadband and IoT business Connexin on a novel speech-to-text platform that integrates sensor networks for assisted living.
Innovate UK, 2023-25, Dethlefs, Cheng
This is an impact acceleration project that enables rapid analysis of text documents from interactive feedback and sparse human labels, by leveraging our previous research in NLP and language models.
HEIF, 2022-23, Dethlefs, Dixon
We propose an industry-compatible advance in accumulated fatigue assessment via novel integration of physical modelling and machine learning.
With the EEI and University of Sheffield.
EPSRC Supergen ORE, 2021-22, Dethlefs
We develop novel algorithms to improve stability testing of new product formulations and accelerate their launch from initial conception and design through to testing and manufacturing. This will accelerate product improvement and enhance Reckitt's capacity to react to customer feedback and product reviews.
Innovate UK, 2021-23, Dethlefs
Working with BIBA at Uni Bremen to develop a novel computer vision approach for in-situ identification of ice on wind turbine blades.
German Federal Ministry of Education and Research, 2021-22, Chatterjee
We are part of the Aura CDT on Offshore Wind Energy and the Environment lead research theme "Big data, sensors and digitalisation".
With partner institutions Sheffield, Durham and Newcastle.
EPSRC, NERC 2019-27, Dethlefs
See DetailsWe apply research in data analytics, NLP and AI to analyse company documents, identify patterns, extract information and generate insights towards more effective workflows and informed decision support.
Innovate UK, 2020-22, Dethlefs, Dixon
Research to address the openness, uncertainty, variability, and interplay of safety and security in multi-robot systems.
See DetailsEU H2020, 2021-24, Dethlefs (CI)
We research hierarchical decomposition approaches for fast learning in natural language generation systems that can produce high-quality outputs from sparse and unseen knowledge graphs.
Diffbot (industry grant), 2017-19, Dethlefs
The aim is to semi-automate the analysis of legal documents with an intelligent software that can extract prominent themes from a document, and identify likely areas for legal attention.
With legal firm rradar.
HEIF, 2019, Dethlefs
Project investigating 3D Analysis of Maxillofacial Growth in Patients with Cleft Lip and Palate. With colleagues in Medical Engineering at Hull.
Wellcome Trust, 2027-19, Ma
We're proud of Dr Lydia Bryan-Smith for passing her viva on the topic of "Using Big Data and AI to Dynamically Predict Flood Risk", forecasting floods from social media and using AI to tackle hydrodynamic simulations. Well done!
Congrats to Sam, who won a competitive Turing Internship Network placement with Everybody Counts Ltd to develop novel algorithms to understand how dyscalculia, dyslexia and forms of autism can affect learning. Very proud of you, Sam, for pioneering in this important area! 🎊
With conference season upon us, Sedat went to network at the Second Workshop on Multimodal AI in Sheffield to present his research on virality prediction on memes from multimodal cues. See his LinkedIn.
Fantastic news: Joyjit's KTP project with Reckitt was covered as a success story on Ashorne Hill's website, the main KTP training provider. Read the article here.
Huge congratulations to Vic for winning a Turing Enrichment Award for her research on memes from the Alan Turning Institute (of only 20 across the UK)! Very proud! (write-up to follow)
We co-lead a new EPSRC Centre for Doctoral Training in Offshore Wind Energy Sustainability and Resilience with Durham, Loughborough and Sheffield. Looking forward to at least 65 new PGR students on cutting-edge solutions to OSW challenges and driving the UK’s transition to net zero.
Some of us are off to NAACL in Mexico City in June to present our BDA entry on Detecting Persuasion in Memes Across Languages. We explore the effects of various sources of external knowledge and did really well on the leaderboard.
It's only 11 days left until our research workshop. This year we celebrate with our own Charity Christmas tree, made by the FabLab in donation to Dove House Hospice (spot it in the picture).
Joyjit spent his Best International KE Initiative Award fund learning about enterprise and business opportunities at the University of Cambridge, the Institute of Continuous Learning (ICL).
Callum worked with Tobisha Europe's Speech Technology Group at Cambridge this summer to explore the feasibility of human-in-the-loop machine learning to classify turbine blade damage from images using dialogue. Paper here.
Nina will be working on a Researcher in Residence grant with the Offshore Renewable Energy (ORE) Catapult to embed our work on natural language for O&M into the industry in the form of a ChatOSW system. One of our smallest yet most exciting grants!
Congratulations to Dr Albert Dulian who successfully passed his viva on Short-term motion prediction of autonomous vehicles in complex environments: A Deep Learning approach. Well done, Albert!
Lydia presented an amazing demo at this year's Humber Science Festival showing how modern AI models represent language in context. You can see different languages represented with an outer space touch...See Lydia's video here.
Nina joined a discussion panel on data sources and analysis for offshore wind at this year's O&M Conference, hosted by Renewable UK and OREC. A very interesting day and lots to think about and explore further.
We're busy this summer with Callum and Eamonn first presenting their Offshore Wind research at WESC in Glasgow, and then at YISEC in Bradford. Victoria presented her work on meme classification at ICMR in Thessaloniki, and Joyjit headed to Alberta, Canada to present this KTP on a funded visit to Amii. Ifeoluwa had her first paper accepted at LSD and is heading to Gothenburg in September. Congrats all!
Joyjit presented his KTP with Reckitt last evening at Uni Hull's 10-year KTP Celebration Event under the motto Team work makes the dream work. Thanks so much for representing BDA in such a positive light! You can read Joyjit's full narrative of the event on LinkedIn.
Happy to announce our new KTP with Hull-based broadband and IoT business Connexin. We will develop a new Smart Speaker system for assisted living, bringing together novel aspects of speech-to-text generation and sensors. We're looking to recruit a KTP Associate (PhD-level) soon...
Congratulations to Onatkut Dagtekin for passing his PhD viva on the topic of Deep Learning for the Early Detection of Harmful Algal Blooms and Improving Water Quality Monitoring. Well done Onatkut and good luck in Essex. Keep in touch!
We're at the Hull Science Festival this year! Erv is demonstrating (hands-on) how AI can be used to generate create music collaboratively and Lydia is joining the Viper team to talk about the role of HPCs for AI.
Joyjit spent a month at the University of Bremen in Germany, collaborating with the BIBA Institute on a project on ice detection on turbine blades. This research stay was funded by the Federal Ministry of Education and Research via the Green Talents Award. Thanks to Maite Alvela and Hannes Gelbhardt for hosting!
Onatkut participated in this year's Climate Change AI Summer School 2022, and presented this very timely research on detecting harmful algal blooms with deep learning at IJCNN in Padua, Italy this summer.
Joyjit was interviewed by IEEE Spectrum on AI's impact on wind and solar energy, stressing the urgency of more uptake, and the potential of natural language as a way to enhance explainability. Nina presented our research on wind and language to Toshiba Europe's Dialogue Systems group. It was lovely to chat with everyone there.
We are proud that our very own Joyjit has been named in Forbes 30 under 30 Europe this year under the Manufacturing and Industry category - people that create the products, methods and materials of tomorrow! Congrats Joyjit, well deserved!
We're part of the Faculty's new Centre of Excellence for Data Science, Artificial Intelligence and Modelling (DAIM). You can join us a taught MSc student, PhD student, or as a new member of academic staff.
Nina is extremely excited to have received the Inspired in Hull Award for "Excellence in Supervision". What an honour seen the amazing colleagues that were nominated, having a brilliant cohort of PhD students like our BDA bunch is a real advantage!
We're thrilled to be hosting a Social session at ICRL 2022 - featuring a panel discussion with a set of experts from the wind and renewables sector, branching across academia and the industry. We hope to explore the opportunities and risks of AI in the transition towards Net Zero. See our dedicated page here.
Read our latest Conversation article! We discuss the carbon footprint of modern AI and how it contributes to furthering the divide between developed and developing countries.
Our (first) NERC discipline hopping Hackathon in AI for Sustainability took place w/c 28 March, jointly organised by Computer Science and the Energy and Environment Institute. We spotten hedges from the air, predicted the structural health of wind turbine monopiles, and predicted floods from social media. See our hackathon page here.
Congratulations to Bobby Houseago for winning the Early Career Researcher Poster Prize at the Supergen Annual Assembly this year for his poster on rapid fatigue assessment in offshore wind farms. Yay!
Thanks all who made it to our annual sparkly Christmas research workshop to present your work, discuss ideas, and eat some cake... Looking forward to next year!
We're massively proud of Joyjit who got selected for a Green Talents Award from amongst 25 researchers from around the globe. The German Ministry of Education and Research awards these prestigious titles + a fully funded research stay in Germany to young people that have made an outstanding contribution to research in sustainability.
The University are part of an amazing initiative to create a hedgehog-friendly campus, see here. This involves raising awareness of hedgehogs and their habitats, creating safe spaces, including bug hotels and hedgehog shelters, and keeping our campus clean and tidy. We're very excited about this!
Congratulations to Dr Joyjit Chatterjee, who has successfully passed his viva, completing his PhD on Explainable AI for O&M of Wind Turbines, see here. Well done Joyjit - and we're excited that you're staying with us in a postdoctoral role.
We won an EPSRC Supergen ORE grant with Agota Mockute at the EEI and Lizzy Cross from Sheffield and a group of strong industry partners. We'll be predicting fatigue levels of wind turbines using deep learning and physical modelling. See details here.
Congratulations to Joyjit for completing the Nudge Global Impact Challenge that selects young research leaders to develop their own sustainability research plans. See Joyjit's promotional video pitching his PhD project here.
Congratulations to Dr Annika Schoene, one of our founding members, who has successfully passed her viva and is leaving us to take up a postdoctoral position at the University of Manchester. Keep in touch, Annika!
We are really excited to be working with Reckitt on a 2-year KTP project to develop novel AI techniques to improve stability testing and new product development, and accelerate product development from design over testing to manufacturing. What an amazing opportunity!
Joyjit and Annika went to NeurIPS-2019 in Vancouver to present their papers Natural Language Generation for Operations and Maintenance in Wind Turbines and Dilated LSTM with ranked units for classification of suicide notes. Well done!
Annika went to Hong Kong to present her paper Dilated LSTM with attention for Classification of suicide notes at the LOUHI workshop at EMNLP-2019. See here.
Joyjit was part of a Uni Hull team to enter the Offshore Wind Hackathon in Glasgow, applying deep learning to predict the power availability in wind farms.
We are looking forward to working with the Spencer Group on a new Innovate UK-funded Knowledge Transfer Partnership (KTP) to integrate artificial intelligence and natural language processing into corporate workflows. Job advert to follow shortly.
We are part of Hull's new EPSRC and NERC-funded CDT in Offshore Wind and the Environment together with partners Durham, Newcastle and Sheffield. See here for website and PhD student recruitment.
Joyjit presented some early PhD work at the Northern Lights Deep Learning Workshop in Tromsø - doing some networking and seeing the Northern Lights. See his presentation and talk here.
We are hosting a 2-day GPU and CUDA programming course right after the Deep Learning Winter School 2019. The aim is to gain an understanding of GPU hardware and a competence in CUDA GPU programming for scientific applications, with a particular focus on deep learning. See here for details and registration.
On 7th December the Big Data Analytics Research group gathered for its annual Poster Workshop and Christmas Lunch.
Joyjit spent some time with the ORE Catapult in Glasgow last month. Read about this visit.
We're excited about our upcoming Christmas Workshop and Lunch on 7 Dec. See our flyer.
We are organising the 2nd Deep Learning Winter School on 22-23 January 2019 as a hands-on introduction to the theory and practice of deep learning, including talks from leading researchers in the field a tour of Viper See here for details and registration.
The new Aura wind energy PhD cluster met for a kick-off meeting with the 4 new students, including Joyjit from the Big Data Analytics group. Read Joyjit's report.
We're recruiting for a post-doctoral research assistant to work on an EPSRC-funded project on Transparent AI. See here. Deadline 15 Nov 2018.
We're recruiting a PhD student on a 3-year fully-funded scholarship on data analytics (particularly deep learning) for water sensing. See here. Deadline 23 Jan 2019.
We're recruiting a PhD student on a 3-year fully-funded scholarship on data analytics from IoT devices for modelling flood risks. See here. Deadline 23 Jan 2019.
Once the start date and funding for my PhD was confirmed, I started searching the internet for what makes a successful PhD student in Computer Science. ...
Read more
David Benoit from Chemistry and Nina Dethlefs gave a talk at the British Science Festival on "Let the AI In".
Read more
Following up on our talk at the British Science Festival, we published an article in The Conversation --- AI: there’s a reason it’s so bad at conversation.
Alex Turner got his EPSRC grant on 'Using Epigenetically-Inspired Connectionist Models to Provide Transparency In The Modelling of Human Visceral Leismaniasis'.
See details.
Today we had "big data" cake to welcome 3 new team members who will be working on deep learning and computational biology.
Read more
After having my paper accepted for the 2018 WISDOM workshop at KDD, I felt both nervous but excited. ...
Read more+44 (0) 1482 465994
Email one of our team
Robert Blackburn Building 3rd floor University of Hull Cottingham Road HU6 7RX, Hull, UK
