Brarista replicates professional bra-fitting by eyesight with personalised sizing recommendations. A B2B software, it is free for the end-consumers to use via partnered retailers and brands. End-user consumers simply and securely upload photos of their upper bodies, wearing a tank top, to find out their correct bra sizes specifically in all the products they are interested in. Barista license its software to a wide range of online and omni channel retailers, to improve bra-wearers' experience, well-being and reduce sizing confusion/anxiety. Currently 90% of returns for retailers are fit and size related enabling retail partners and B2B clients to significantly reduce the cost and impact of returns to their business.

TOffeeAM is a design for additive manufacturing software based on topology optimisation (TO). The platform automatically builds the optimal design of an engineering component. Using artificial intelligence the system can be used to improve the accuracy of the predictors during the optimisation phase and to speed up the design time.

NeuCurrent is a one stop data platform for omni channel retailers. NeuCurrent uses AI, automation and best industry marketing expertise to allow marketers to create personalised recommendations and customer engagements over the entire customer journey without fuss.

An AI platform that helps data scientists identify algorithmic bias related issues in their models pre-production and gives them options of different repair pipelines.

The selection of treatments for multiple sclerosis is subjective, with trial and error that costs hundreds of millions every year for the NHS. The Queen Square Analytics solution is a machine learning (ML) enabled algorithm that applies on brain images to divide patients into subgroups and quantify the risk of disease progression and select the right treatments in an objective way.

Machine learning and AI solutions available on portable camera platforms, such as camera traps, to monitor animal biodiversity.

"AI solutions for the retail sector, including a product search tool which enables customers to find the products they are looking for as quickly as possible, using images rather than keywords

Cadeera delivers three primary capabilities: 1) Product search and discovery through an intuitive search experience, enabling customers to find the products they are looking for as quickly as possible, using images rather than keywords and text-based filters. 2) The ‘Deja-vu’ engine provides a module which allows the user to take a picture and find products as close as possible in the inventory. 3) Data quality, a module that addresses one of the most fundamental issues clients are experiencing around data entry, labelling errors and missing values.

DynAlkon helps consumers to understand their preference and desired style through visual clues, effectively leveraging imagery retailers already have and making these images shoppable. It applies state-of-the-art computer vision and machine learning algorithms to analyse product images, description and user interactions. The technology complements and can easily be integrated into existing eCommerce solutions, immediately enhancing the shopping experience.
"

Kavida is a supply chain digital twin platform that enables enterprises to make intelligent resiliency decisions. Using artificial intelligence, the system detects threats before they occur or in real-time. After detection, it simulates to quantify the financial and customer impact and aids decisions to mitigate the impact.

As a team of artificial intelligence, privacy and product experts, Bitfount understands the hurdles to finding and incorporating the right data into analytics, applications and machine learning models. Bitfount has created a collaboration tool allowing data providers, data scientists and governance teams to work together without compromising on privacy or quality.

A hybrid physics and AI approach for optimising industrial process control in real time.

Given all material has value, waste should not exist – simply, materials are in the wrong place. The Recycleye team is working towards making this a reality by disrupting the waste management industry with an ultra-low-cost, rapidly deployable, decentralized, scalable, digital & fully-automated robotic sorting solution – the Recycleye mini-MRF (mini material recovery facility). Recycleye has already developed a state-of-the-art computer vision system using recent advances in deep learning and AI capable of classifying items by material type and brand. The company is developing other technologies that will shape tomorrow’s waste management infrastructure and accelerate the world’s transition towards a circular economy.

Statmetrix is a sports technology platform that enables more efficient identification and development of sports talent using video. Globally, there are over 25 million youth football players; 3.35 million of them in the UK, playing at the grassroots level. The problem is that providing performance analysis needed to help the players improve is a manual time-consuming process for coaches. As more players sign up hoping to become professional, many do not get the feedback and evaluation they need. Meanwhile, the technology available to established top division clubs are expensive and inaccessible for grassroots teams. Statmetrix offers a cost-effective video solution that uses artificial intelligence to automate previously manual performance analysis processes thereby enabling coaches to coach at scale.

action.ai empowers its clients to launch next generation Chatbots quickly and cost effectively. Existing Chatbots are failing because they do not understand users’ everyday language. action.ai’s industry-leading technology processes human language to provide sophisticated understanding, no matter how people express themselves. The company’s technology works via voice and text, and can be offered via any channel e.g. via the Web, smart speakers, or Instant Messenger. Next generation Chatbots provide intuitive and frictionless interfaces whether to improve sales conversion or to greatly reduce the costs of customer service. Clients own and control their Chatbots, including all customer data. action.ai is one of the most highly awarded AI companies and its technology is used in sensitive commercial environments, for example, in banking.

Product design engineers everywhere are presented with more ambitious innovation briefs, while subjected to more limits on cost, manufacturing, sustainability, and time. Simultaneously, new technology and materials are emerging – leading to missed opportunities. It has developed the first design software for Additive Manufacturing that can evaluate and recommend multi-material allocation within a 3D object in conjunction with topology – so every option available to designers is explored.

Adoreboard, a Queen’s University Belfast spin-off, is a pioneer in Human Experience (HX) using Emotion AI, the next frontier of artificial intelligence. Adoreboard unites the customer and employee experience through its Human Experience (HX) measurement platform Emotics. Adoreboard’s mission is to bring emotionally-intelligent insights into the digital world with its emotion recognition technology that measures Human Experiences that span customer and employee interactions across the enterprise.

Adoreboard is a Gartner Cool Vendor in AI for Customer Analytics. Adoreboard in it’s pioneering work on Emotion AI will avail of access to world leading infrastructure, expertise and networking opportunities with one of the industry leading programmes in the advancement of artificial intelligence.

Continuum Industries supercharges design of linear infrastructure through automation and optimisation.

How has Machine Intelligence Garage helped the company?
Continuum Industries applied to Machine Intelligence Garage to increase its capacity to do testing of its optimisation engine and to reduce costs of deployment to initial customers.

illumr helps organisations better understand and predict patterns of behaviour that affect them by providing ‘Explainable AI’ solutions. This helps enterprises better understand the problems they are facing rather than making predictions via black-box solutions. Gaining understandable insights from data used to be a time and cost intensive exercise, often confounded by the human bias of the analysts involved. illumr solves these issues by building upon decades’ worth of research. It has two key products that can not only remove the inherent bias in data, but also explain – in human understandable terms – the natural patterns in data; these hypothesis-free insights are usually missed by all other tools and methodologies.

Plyable exists to catalyse the digital manufacturing revolution. Its mission is to build the definitive global marketplace for composite tooling.

Pukket is a customer advocacy management platform that enables brands to create word-of-mouth at scale on social media. It does this by encouraging its customer base to start sharing their experiences on social media platforms. It uses computer vision and NLP techniques to automatically detect, evaluate and reward posts in real time.
Pukket’s artificial intelligence capability gives it an advantage over other competitors in the multi-billion dollar industry. Enhancing such capabilities and making sure it does so ethically will be its main priorities over the coming years.

Sermaurei is a clinician-designed audio electronic health record system. The web application records, stores and transcribes clinical conversations between a clinician and patient to enable clinicians to focus on their patients without worrying about accurate documentation.

Sermaurei protects clinicians & patients. Its mission is to create consistent real-time accurate, transparent and intelligent records (for patients & clinicians) with outcome based health data to improve healthcare delivery. It aims to be the gold standard in clinical record keeping and make the process less burdensome with the use of NLP, speech-to-text and machine learning. Sermaurei applied to Machine Intelligence Garage with the hope that even if it were to be unsuccessful it would learn a lot from the application process. Should it succeed it knew that being on the programme would provide it with invaluable access to high performance computers and introductions to artificial intelligence experts and services that will help it develop, refine and scale its product better and faster than we would be able to do without Machine Intelligence Garage.

Orthodontics is democratising access to dental health & orthodontics globally through its technology platform, team of digital dentists, customised products & smart devices. It is a $950bn global digitisation opportunity.

Orthodontics is essentially a computer vision problem. It processes a lot of data in multi-view images and is building algorithms to augment the capabilities of the digital dentists that work on our platform.

Alchera Technologies is an AI software startup providing the intelligence layer that enables road infrastructure to be operated and optimised dynamically, in real-time. Its product is Alpha: cloud AI software for use in enterprise-grade intelligent mobility and intelligent infrastructure applications, providing millisecond-latency, accurate, auditable mobility data, at very large (country/city) scale.

The programme has connected Braendler Engineering with the AI/machine learning community in the UK, and it can see where the various workshops and seminars will be invaluable to it as it brings on investment and scales the business. Furthermore the access to the computer power will accelerate the work that it is doing in machine learning, which will enable it to commercialize this earlier than would otherwise be the case. It is very much looking forward to working with Digital Catapult through the Machine Intelligence Garage programme over the coming 12 months.

DeepMiner is an AI-driven B2B search engine that allows users to spot and address emerging threats and opportunities to their business or organisation. With an ever-increasing stream of information, and business disruption increasing at ever faster rates, it is critical for those charged with the strategic direction of an organisation to know the very latest information and trends in markets, competitors and new opportunities. DeepMiner’s search engine allows users to search global news sites, social media and public data sets to find up-to-date and relevant information without the added bias and optimised search results from current alternative search engines. This unfiltered search helps give leaders the true picture of the world around them.

Ducit.ai is an artificial intelligence and machine learning open banking platform which aims to democratise banking data for the benefit of consumers and small businesses. Its mission is simple – to deliver intelligent banking for all. Its vision is to do it through the power of artificial intelligence and machine learning. Its aim is to accelerate the impact of open banking and deliver its benefits faster to consumers and small business. Its Account Information (AISP) APIs offer artificial intelligence and machine learning powered products for personal finance management, financial affordability and fraud detection for the banking industry. Ducit.ai’s Payment Initiation (PISP) API allows customers to initiate payments and transfers securely, conveniently and at a significantly lower cost than the incumbents.

The High Street in the UK is struggling. Margins are low, stress is high. catsAi’s first step along the high street is with bakers. That industry currently generates waste of circa 25% of production (National Bakers Association). It believes that if it can help make better predictions, it can reduce waste, and thus increase margins. Machine learning is a toolset that it’s deployed into these small businesses. catsAi’s ‘brain’ learns the patterns of each individual store and product list; resulting in better sales predictions for the coming week, and thus up to 75% reductions in waste.

Manas AI specialises in two areas. Its first product automates and accelerates high volume real time ‘what next’ decision making. This is used to address a range of challenges – from intraday trading strategies to predicting customer behaviour to understanding disease progression. Its second product automates the extraction of meaningful insights from communication. This is used to address the challenge of making sense from the increasing volume of customer communications.

Panakeia is building universal engine for one-stop precision cancer diagnosis.

How has Machine Intelligence Garage helped the company?
It is looking to get assistance on machine learning infrastructure and expertise as well as data governance and ethics.

Machine Intelligence Garage has helped Panakeia to think about the future direction of its ethics and data governance strategy.

Techspert.io enables businesses to make better, more informed decisions. It uses AI driven mapping of expertise and its attribution to connect businesses to the right experts. Its platform provides efficient knowledge exchange service to consultancies, private equity firms and corporates.

Neura Technologies was started from academia to tackle real world business problems by bringing state-of-the-art research to your front door. Its expertise in AI and machine learning means all you need to do is supply data and watch as it transforms into fresh insights and actionable intelligence. It has the tools to assist businesses from start to finish on a global scale, from data collection, to analysis, to deployment – a true one stop AI shop.

Greyparrot use computer vision and deep learning to solve industry challenges and create positive social impact. Drawing on the latest research and providing turnkey computer vision solutions to companies looking to partner for their AI requirements involving image or video datasets. Image and video recognition APIs is offered which can be integrated into systems. The waste recycling API enables sorting of waste types to automate and improve recycling processes. Another example is car API which recognises 6,500+ car types and can be used to streamline insurance quotes, give second-hand price comparisons, tag images for e-commerce and more. Greyparrot is committed to implementing AI in a practical, scalable and ethical way.

KageNova is developing AI technologies to engineer interactive realism for next-generation immersive experiences, such as virtual and mixed reality.

Vision AI for scientific understanding. V7 gives cameras in scientific laboratories artificial intelligence to visually understand experimentation for automatic documentation of protocols, saving up to half the time required for a procedure. Its training platform – Darwin, allows its technology to scale to multiple areas of science and beyond.

Remo.ai is a platform for open and reproducible Computer Vision. AI shouldn’t be owned by any individual company, as such further research into promoting open access to AI research is being developed, making the field transparent and accessible at the same time. For researchers, this means a platform to host and benchmark a live version of their models. For practitioners and developers, it means a place where collaboration on projects and easy access to AI research and datasets is possible. For companies, it’s a platform to easily build applications of the latest research.

Prodo.AI make machines write code for humans.

MAIO, a smart industry solution developed by Smartia Ltd, is an AI and Data technology framework that lets companies connect to industrial assets and collect data to help organisations find insights and value from their operations. MAIO helps companies improve efficiency, quality and productivity, reduce costs and waste and increase profitability and competitiveness.

Antimicrobial resistance (AMR) is a growing concern with a projected global mortality rate of 10 million per year by 2050, costs exceeding climate change and catastrophic health impact. Antibiotic overuse is directly correlated with AMR, yet 20-50% of antibiotics prescribed are unnecessary or inappropriate. Abtrace uses artificial intelligence techniques to extract clinical meaning from healthcare notes building a next generation decision support tool to optimise antibiotic prescribing. The outputs can transform what clinical decisions are understood, measured and tracked. The application of this technology will in turn reduce AMR rates, improve health outcomes and reduce wider societal costs.

By combining the most advanced technology, data driven knowledge and its expertise in musculoskeletal disease, myrecovery brings together patients, healthcare providers, academia, life science companies and regulators. myrecovery crafts tools which provide the entire sector with richer data, deeper insights, clearer decisions, smarter discoveries and better outcomes. It is dedicated to improving the lives of more people with musculoskeletal disease by evolving their care and treatment.

Machines With Vision is building location-as-a-service for digital rail, autonomous driving, and delivery robotics. It does this with a precise, reliable and low cost approach based on scanning and matching a unique ground surface fingerprint. This approach addresses failure cases in existing technologies, such as GPS in urban or underground situations, and enables further use cases in last meter transport automation. Machines With Vision’s customers are national rail maintenance divisions, train operating companies, tier-1 automotive testing, and logistic/mobile robotic suppliers. The underlying technology is based on neuromorphic/event cameras, requiring it to develop a range of new algorithms for interpreting and processing such fast and novel data.

CollectiWise’s cloud software-as-a-service helps teams collectively make complex assessments and estimations as to their resource needs for projects and time needed to finish.

ZEG.ai has built a world-first 3D AI that allows auto-generation of 3D models for e-commerce, XR, and games. Historically, 3D model generation has been manual, time consuming, and prohibitively expensive. ZEG.ai’s 3D AI allows for model generation through intuitive approaches such as 1-2 images, text, and/or voice.

UrbanTide’s award winning USMART software unlocks the power of data to help cities and communities work better and smarter. USMART provides the entire data pipeline from data ingestion to insight exploration with advanced data discovery. Its artificial intelligence applications in USMART are making smarter places unlocking revenue opportunities for cities and communities.

Cyanapse develops software tools to improve the perception of images and data, using methods based on AI and neuroscience research. The company provides image editing solutions that enable users to automate their image editing tasks and to enhance their visuals using photorealistic filters. The platform helps users identify the best transformations for their visuals to communicate their messages better.

Data Language is an expert in data driven product innovation. Its cross-sector data science consulting practice excels in the design and implementation of AI and machine learning solutions that solve real business problems. It is experienced in designing machine and deep learning models for the analysis of data and media at scale. Its Tagmatic product suite uses ML to solve media automation challenges for the news and publishing industry.

Disruption within the food supply chain is being driven by macro trends, including the exponential growth in world population, which coupled with increased protein consumption have put pressure on existing livestock protein sources. Legislation limiting emissions from livestock means that producers cannot simply farm more animals to feed the increasing demand, but instead need to become more efficient to get more from less. CattleEye is a video analytics platform powered by deep learning which solves this problem by providing insights to livestock farmers and the wider supply chain which increases farm efficiencies and improves animal welfare.

Headlight AI is a sensor technology company that is making sensors smarter. Its adaptive sensor processing software makes individual sensors more intelligent and robust. Its proprietary models train sensors to adapt their operational parameters in real time and depending on their surrounding environment. This enables sensor systems to operate in confined spaces with poor visibility and other extreme conditions in which they typically fail, such as: rain, fog, snow, dust or smoke. Headlight AI aims to sense where no one else can sense, process data and offer insights faster than anyone else can, and shine a light on problems people didn't even know existed, and offer solutions.

Contilio is a London-based deep tech startup solving one of the biggest challenges for the $12trillion construction industry. It uses 3D computer vision, deep learning and BIM to automate the broken performance reporting process for mid-size and large construction projects. Its world first analytics platform turns site data into real time insights on project progress, performance KPIs and risks. This enables project owners, contractors and investors to make timely and accurate decisions and realize significant productivity, cashflow and compliance improvements.

AWAYTOMARS supports companies with quality creative information using co-creative tools and public engagement. It provides this with an AI based software that will co-create with designers to achieve better products.

Lobster is an AI powered platform where creatives and brands can find and license user-generated content (UGC) from users on 10 platforms – Instagram, Facebook, Flickr, YouTube, Vimeo, Google photos, Verizon Cloud, Foursquare, Dropbox, Vk etc. Its technology curates and ranks the images that budding creators list for sale. Creatives looking for fresh, authentic, off-the-beaten-track visuals can take advantage of Lobster’s AI/ML features such as Super Resolution (up to 4x higher res for lower res files), Style Matcher (search by style), Style Transfer in order to reach the best image and instantly obtain a royalty-free license with a smart contract.

Hoxton Analytics provides highly-accurate footfall, gender, occupancy and directional flow data, based only on observing visitor footsteps. The company never collect any personally identifiable information and are fully compliant with GDPR. With each new deployment, the size of its datasets (labelled and unlabelled) grows ever faster and this enables the company to train ever more complex neural networks.

The next generation in audio identification/recognition; a solution that can recognise every performance of every song everywhere, on every stage and on every platform – live, online and offline. Breathe Music will be accurate, ubiquitous music recognition for fair music royalties – live and online.

ChAI forecasts changes in commodity prices over time horizons that are useful to those who have responsibility for the risks associated with these costs – e.g. purchasing managers and financial controllers. The service is able to articulate the confidence it has in its predictions when they are made. This means that the user can incorporate their own sensitivity to a movement when making a judgement based on the system’s outputs.

Media Localisation is expensive, time consuming and of poor quality. Growth of streaming services has accelerated both consumption of localised content and demand for improving its quality & speed of production. However, innovation in localisation services have not kept pace with the needs of the entertainment industry.

DeepReel leverages the power of artificial intelligence to create the next generation of media localisation services for the entertainment industry.

DeepReel localises video content into different languages with flawless synchronisation of actors' lip movements to the new language and offers clients the ability to choose different actor voices.

NextQuestion delivers effective data cleaning for retailers using AI. The company provides solutions for supply chain inefficiencies in the form of reduced waste and out of stocks. It is an algorithm-as-a-service company, with the latest version of its solution estimated to instantly deliver 0.2% revenue savings to clients and a value proposition which is constantly improving due to the self-optimising nature of its models.

EnVsion helps creators save time & produce more engaging content at scale by automating their video workflows using our cognitive AI technology. Its end-to-end solution is used to improve workflows from video enrichment & discovery to video editing & creation.

Loomi.ai is the first AI platform that enables professional users to create and train personal AI assistants to help organise and prioritise their online world. The company utilises the existing online data of users such as email, social media or corporate accounts to securely train and personalise each user’s AI assistant. For individuals, the assistant helps generate ToDos, manage email and calendars, research and fact-check information and much more. For businesses, PAs can be enabled for employees, customers or partners that can organize information and focus on the key tasks and insights which significantly improve productivity.

Figaro (by FeedForward) is a B2B search and metadata engine for music catalogues, powered by machine learning. Figaro learns the relationship between the words we use to describe music and the music itself in order to offer scalable metadata management and intuitive music search.

There are two parts to the existing service:

1) Metadata Management - Figaro enables music catalogues to manage their descriptive metadata (such as genre, mood, and instrumentation) at scale. Figaro learns a catalogue's taxonomy, capturing the custom meaning of descriptive words. It becomes a helpful assistant for metadata managers by automatically predicting metadata from their taxonomy for tracks, enabling the catalogue to make new music searchable quickly and achieve consistency across their metadata.

2) Intuitive Search - Figaro similarity search enables users to weight search results based upon elements of the music that are important to them, such as genre, mood and instrumentation. Intuitive search features are being expanded to take into account other search inputs and the individual intent of each search request.

Figaro can be accessed via API or app.

Scholarcy helps clients add intelligence to their content. It saw a need for a product that could read any document and give background reading, key facts and an overall summary. It developed Scholarcy, initially aimed at the edtech sector, to help researchers, students and journalists read and understand the pile of papers and reports on their hard drives or on the web. While there are plenty of tools that help you find papers, Scholarcy is the only tool that will actually read a full paper or report for you.

DeepZen works with sector specific solutions that will eliminate the need to use studios and lengthy recording sessions by creating human-like quality of speech with emotions that can be edited by using intuitive tools. Users can pick voices from an existing library, modify and tweak them, then generate voices directly from texts.

Humain’s mission is to make digital human creation easier for everyone by automating its labour-intensive processes.

Humain will use its deep understanding and expertise of creating digital humans and combine it with advances they have made in machine learning to deliver high-quality 3D character assets at lower cost and in a shorter timeframe to AAA video game developers.

Chirp is a technology company addressing the need for frictionless, portable and secure device-to-device communication.

Its core data-over-sound technology provides a robust means of encoding, transmitting, and decoding digital assets, with maximum device interoperability from the latest smart hubs and IoT devices to legacy mobile phones and low-cost toys. By using sound, it is able to provide connectivity solutions that are truly device agnostic, do not require device pairing, work offline, leverage existing audio infrastructure such as streaming and public address systems, and can be used for both one-to-one and one-to-many communications. It offers a broad suite of SDKs for developers to seamlessly integrate its technology, covering a range of applications including authorising user credentials, ticketing, payment transactions, and onboarding smart devices.

As the world becomes more connected and digital the demand for digital 3D assets grows every year, yet they remain incredibly expensive to create. Their production is fundamentally unscalable, relying almost solely on artists sitting in front of screens for hours on end. These digital 3D artists are highly skilled and expensive, and much of their output must be redone or changed at great expense as the design brief evolves. For massive game worlds the asset costs can reach into the hundreds of millions.

Kaedim’s product is AI powered software which produces 3D meshes from 2D input images in 1/100th the time it would take a human designer.

The core AI model can generate multiple 3D models from a single input image. This is combined with a highly defensible styling module which can produce 3D models in a specific studio or project style. To do this Kaedim trains the model on data from its customers, enabling it to learn and replicate their aesthetic. By selling licenses to firms with large archives of data it becomes much harder for competition to replicate its results as they will not have access to this crucial resource.

Glass use AI technology to interpret open text on the web.

Glass’ mission is to read the entire internet to build an unparalleled open resource for the world’s researchers. Most organisations have to do research in order to understand markets and find new customers. With this innovation in the field of machine language understanding, Glass is building a new capability for market, economic and social research. Initial users are data scientists, economists and market analysts. Through Glass, users can gain insights into themes and sectors, monitor trends, find opportunities, and perform market analysis. The application is built entirely with AI that digests content from the web, the world’s biggest research resource and system of record.

Machine learning and computer vision have an immense potential to revolutionise industries, and the media in particular, yet it remains incredibly long and costly to build visual AI applications. The most important driver of that complexity is data, and the difficulty to manage and curate datasets at the kind of scale required by AI.

Metalynx builds data curation tools for computer vision teams, helping them accelerate development by removing much of the painful data visualisation and curation work. They help machine learning teams understand their datasets through intuitive data interfaces and automated methods to identify the right visual data for their application.

axial3D makes 3D printing in medicine routine.

Complex injuries are hard to interpret in 2D medical images; however, 3D printed models of these images have shown to have a huge impact on a patients care. The process of turning these images into 3D data for 3D printing is extremely complex, time consuming & expensive, obstructing worldwide adoption. axial3D is creating a unique software category in this early stage market, utilising machine learning to automate every process in creating functional 3D models from 2D files and eliminating costly human intervention. This opens the multibillion dollar medical 3D printing market to the masses.

Charisma.ai creates interactive entertainment to bridge the gap between TV series. Its focus is to fill the vacuum between series, where viewers are missing the characters and stories they love, with deeply engaging and immersive experiences.

Charisma.ai brings characters from the show to life so that audiences can talk to them through their screens, get involved in the story, and change the narrative. Its solution creates interactive episodes which break new ground in storytelling, and allow audiences to be closer than ever before to the stories and characters they know and love.

Yossarian looks to augment human creativity by using machine learning to find novel lateral associations between content.

Current search and discovery tools optimize for literal associations, which work very well for most people’s day to day search activities. However, for individuals working in creative fields these types of returns can be problematic in that they reinforce existing connections and knowledge. Through metaphorical approaches, Yossarian’s tools help individuals to find concepts that they would be unable to discover using traditional search tools, helping them generate new insight and creative direction.

For content providers, capturing – and retaining – viewers’ attention is harder than ever. More immersive and interactive content normally wins, partly why film is losing out to video games.

InPerson AI allows any video media, whether 2D or 360, to be transformed into an ultra-immersive, highly interactive experience with six degrees of freedom viewing.

WILD AI pioneers “dataology”: understanding your body, through your data, and acting on it – to never be ill again.

WILD AI believes the traditional reactive medical model is outdated – you get sick, see the doctor, take drugs and eventually get cured is outdated. It helps you live smarter – starting with runners, cyclists and swimmers, helping them train smart.

Current motion capture solutions are either expensive and difficult to use and scale, or inaccurate and slow.

Zeta Motion solves this problem by revolutionising motion capture and 6D tracking with a unique AI-based approach, providing products which are fast, accurate, affordable and secure.

Parallel AI is the world’s first machine learning as a service (MLaaS) platform and marketplace. Its aim is to make AI adoption cost effective and integration seamless via an accessible marketplace allowing companies of all sizes to gain equitable benefits.

Humanising Autonomy is developing a culture and context specific platform for pedestrian intent prediction to improve autonomous vehicles’ decision making, safety, societal acceptance and deployment in urban environments. It envisions a world of natural interactions between people and autonomous systems.

Wiggly connect patients to clinicians and other health services.

Wiggly design and manufacture touch screens that sit in doctors’ surgeries to consulting online. Wiggly has been providing hosted services for over a decade. One of its strengths is the integration it has achieved with the major clinical systems in use, breaking open data silos and allowing patients to use and contribute to their records more effectively. Wiggly work in an environment where data security and information governance sit alongside clinical safety as key considerations in everything it does.

GeekTalent develops machine learning solutions for a detailed labour market analysis to support strategic funding decisions for recruitment, education and skills.

Using state-of-the-art natural language processing techniques GeekTalent extract and organise information from documents such as CVs and job advertisements and help customers to improve their workflows and results.

Skin Analytics’ provides an AI screening service, when used in primary care; it can improve patient outcomes and significantly reduce the number of referrals and the cost of finding melanoma.

Melanoma cases have quadrupled in last 40 years. A 2014 Kings Fund report stated that there are two million GP appointments for skin cancer and demand is increasing, putting increasing pressure on meeting the two week wait guidance. Skin Analytics’ work with Royal Free Hospital suggested two-thirds of referred patients could have been dealt with in primary care.

Intellisense.io uses artificial intelligence and industrial IoT technologies to develop optimisation software for the mining industry.

Intellisense.io’s out of the box applications have proven to drastically reduce operating costs and improve efficiencies across the extraction, processing and utility systems in world leading mining companies. This is achieved with zero capital expenditure and minimal disruption to mines.

Predina uses AI to dynamically predict and prevent the risk of road accidents in real time. Its machine learning model processes live contextual information, for example, on road conditions, traffic and weather, to infer potential hazards and suggest alternative routes. Predina has built its tech for fleets, insurers, cities and autonomous vehicles.

GTN is transforming drug discovery through interdisciplinary innovation.

Bringing a single new drug to the market costs $2.9 billion and can take up to 15 years, with a high chance of failure. GTN combine machine learning and quantum physics and use its unique patented technology, Generative Tensorial Networks, to discover innovative new drug-like molecules. This will create substantial efficiencies in the whole drug development cycle.

Bloomsbury AI deploy machine learning, in particular cutting-edge natural language processing, to extract information from text documents and develop a question-answering system. It's tool Cape is available to use through an online demo.

How has Machine Intelligence Garage helped the company?
As part of it's involvement in the Machine Intelligence Garage programme, Bloomsbury AI had dedicated access to an NVIDIA DGX-1 deep learning computer. Using this technology they could train algorithms much faster than before whilst creating a better model for text understanding. Guillaume Bouchard, CEO Bloomsbury AI reported that: “Machine Intelligence Garage as a programme is more focused on what resources the startups need. The system made a huge change to the team’s productivity and throughput of our solutions.” Shortly after their involvement in the programme, Bloomsbury AI were acquired by Facebook to further natural language processing research and applications, as covered in TechCrunch.

Entale is an interactive podcast platform that uses a combination of AI and manual curation to generate engaging content for listeners and unique monetization opportunities for publishers. The Entale app makes it possible to indulge your curiosity while listening to podcasts, by adding pictures, links, maps and other relevant content for you to engage with, save for later or use as a jumping off point to explore topics further. Our mission is to bring this high-quality immersive experience to spoken word audio in order to transform engagement and discovery. To achieve these goals, we are automating the detection of relevant content using natural language processing.

Luffy AI is developing a next-generation AI technology that expands the horizon of AI beyond data science and compute-dense infrastructure. Our neural network technology offers significantly higher "Capability per Watt" than current AI frameworks – we can pack more intelligence into smaller systems. In addition, our networks have been designed to include neuroplasticity. This allows us to build networks with the ability to adapt while deployed, without needing retraining. Our framework is designed to target industrial optimisation, autonomous robotic systems and IoT applications. However, it is a general technology that can be applied in any area current AI technologies are deployed.

Maaind is a neurotechnology and AI company building human-centered and trustable AI for understanding and improving the human mind. We have built Aurora, the world's first personalized AI assistant for the mind, helping individuals to better understand, manage and improve their mental wellbeing and mental performance. Aurora provides detailed and accurate insights into stress, emotions, attention and energy levels conveniently and affordably, through smartphones - and soon other platforms. Aurora is powered by Maaind's bio-XLP technology, a multi-modal approach to predicting multiple aspects of mental states accurately, using voice analysis, biometrics, brain computer interfaces and behavioral data. We are on a mission to help individuals make the best use of their minds in whatever environment, task or situation they are in.

Using computers to fix climate change. Open Climate Fix is a new non-profit research and development lab, totally focused on reducing greenhouse gas emissions as rapidly as possible. Every part of the organisation is designed to maximise climate impact, such as our open and collaborative approach, our rapid prototyping, and our attention on finding scalable & practical solutions. By using an open-source approach, we can draw upon a much larger pool of knowledge and skills than any individual company, so combining existing islands of knowledge and accelerating progress.

Synth is the fastest way of making synthetic database for internal development and testing or for exporting anonymised copies of your data to third parties. Users simply point and click at the database and Synth takes care of the rest. Synth uses machine learning to understand your schema and its relations, classify data and entities, and generate a synthetic clone to specifications.

Sundance’s team is composed of multi-disciplinary and high qualified staff that delivers innovative solutions, specially designed, for designing innovative embedded hardware platforms using state-of-art-technologies for accelerating innovative artificial Intelligence algorithms. Sundance’s solutions include highly efficient and low power consumption embedded platforms optimised for running deep learning algorithms used for performing autonomous navigation and for other general computer vision applications. Such solutions are fundamental for the new generation of robotics and autonomous systems.

Alloxentric provide a platform for voice and text bots automating user interactions

Aflorithmic Labs helps brands and enterprises become more relevant to its audience by turning generic content into hyper-personalised audio products. It does speech synthesis which requires heavy, computationally intensive AI modelling workload.

"Alchemy Machines develops leading-edge voice automation software using deep learning and Natural Language Processing. Although the technology is sector-agnostic, it is focussing on the legal vertical initially before expanding into other sectors. Lawyers of all levels of seniority are having to take their own meeting notes, diverting their attention away from high-value fee-earning activities. It aims to reduce this administrative burden by creating customised speech-to-text products which integrate legal lexicons and embed data-anonymisation technology. This enables accurate and secure sector-specific transcriptions compliant with legal frameworks and organisational policies.

How has Machine Intelligence Garage helped the company?
Alchemy Machines first came across Machine Intelligence Garage via one of Digital Catapult’s events where it picked up an initial report by Libby Kinsey. It is aware that previous startups have benefitted from being part of Machine Intelligence Garage and applied as it wanted to be a part of the machine learning community, access support and mentoring."

"Antibody drugs are powerful and effective, but discovery can be slow and error-prone and existing methods cannot work with certain valuable targets. Antiverse is developing a platform for discovering better antibody drugs faster, as well as against these undruggable targets, ultimately providing better medicines for all.

How has Machine Intelligence Garage helped the company?
At this pre-revenue stage with lots of model development, the cost of computer power is a limiting factor; Machine Intelligence Garage’s offering of cloud credits removes that barrier."

"Property owners and operators do not want to suffer property damages and business interruptions. Insurers do not want to receive claims for losses. Balkerne’s platform provides comprehensive and live property specific risk information to empower property owners and operators to understand risk and prompt the taking of appropriate action to mitigate it, both in the long term and immediate contexts. The platform enables the reduction of the scale and frequency of damages, interruptions and insurance claims, whilst also enhancing customer relationships and loyalty. The first commercial customers are large retailers with a pan-UK property portfolio and an internal insurance captive. The subsequent market being pursued is that of property insurers who need to reduce losses and improve customer engagement.

How has Machine Intelligence Garage helped the company?
Balkerne’s platform requires the processing and analysis of huge amounts of data, everything from news, social media, IoT device outputs, Government and geospatial data. Fundamental to our competitive advantage is the ability to process at speed and with both high contextual relevance and geospatial accuracy. Practically, the only way this can be done is through the use of Machine Intelligence. It has identified that the scope of the assistance offered through the Machine Intelligence Garage would be exceptionally advantageous to its technical ambitions."

"CityMaaS is a B2B Machine Learning data platform providing real time and personalised accessibility information and mobility services. It helps businesses to be inclusive of people with disabilities to increase sales and improve customer services. As a result of CityMaaS Assist App, it helps people with disabilities to achieve efficient travel and live independently.

Machine Intelligence Garage has been a programme that suitably provides not only practical resources such as commutation power, but also has the relevant expertise in AI ethics and machine learning, which the CityMaaS team can leverage to scale its technology and product development in order to commercialise its technology and create the impact its mission has set out to do. The programme has provided an initial AI ethics consultation, which provides a high level guidance and expertise for the CityMaaS team to enhance its current theory of change. It’s GCP credit will help CityMaaS team to lower the product development cost from computation power prospective. It is looking forward to working with the Machine Intelligence Garage team further to accelerate its product build and commercialisation."

"Ectropy Technologies Ltd is a London based fintech company focused on providing an end to end digital platform for the global asset management industry. Exponential growth in data and regulatory changes are making it harder for companies to deliver value to their clients. Its platform enables companies to collect, process and forecast data at scale, in a repeatable and transparent way. It also removes behavioural biases, response delays and availability biases.
Ectropy Technologies sees Machine Intelligence Garage as a partner who can act as a sounding board, monitor and incentivise progress, provide technical expertise and data processing support. In addition, having access to compute power would allow the company to accelerate testing of its modules and features in ETL’s architecture. While it is still early days, Machine Intelligence Garage has helped the company with a deep dive into ethical considerations of its service, as well as very constructive ideas on how to operate during challenging times."

Lab 1 provides businesses with data and insights from hidden and difficult to reach communities from emerging social media platforms to the dark web. It services a range of different use cases ranging from trend forecasting to threat intelligence. Machine Intelligence will help Lab 1 access cloud software credits and provide help navigating challenging ethical areas that a data focused software company faces in 2020, as well as making connections with other members of the cohort. Lab 1 has already had a conversation with ethics advisors which has kick started its thinking on ethical approaches to data collection and GDPR.

Machine Medicine aims to accelerate understanding of the brain and the development of neuro-therapeutics. It does this by providing affordable, convenient and interpretable tools for the measurement of behaviour. Our first product is a computer vision platform for the assessment of Parkinson’s disease. Machine Medicine applied to Machine Intelligence Garage for mentorship and access to computational resources.

Mobilized Construction provides real-time data and machine-learning/AI analytics to monitor and forecast road network road deterioration. Its insights enable more effective and lower cost repairs, making roads safer and less congested for citizens. Roads are the foundations of cities, enabling mobility and transportation – yet time relevant, longitudinal, city-wide technical data has never existed to inform local authorities of how hazards are forming and where they on the road network. If you don’t know where the problem is, how do you fix it?

Mobilized Construction has completed paid projects with Transport for London, Northern Ireland Department for Infrastructure, and Bristo City Council and won numerous global awards for its work on roads. Mobilized Construction deeply believes in launching businesses in rural areas to spark rural revitalization. The main trade off is limited mentors and peer companies to learn from and grow with. It applied to the Machine Intelligence Garage because it was really drawn to the yearlong cohort where it can build more meaningful relationships and also learn how other organisations are managing growth, hiring, and product-market fit.

For professional investors who struggle to detect actionable signals from vast streams of news, Premind AI computationally predicts market anomalies caused by mass and social media. Unlike existing alternative data providers, Premind AI uses unique datasets and proprietary algorithms to understand narratives’ market impact, and in turn offers professional investors signals to outperform the market, reduce investment risks and optimise trade execution.

How has Machine Intelligence Garage helped the company?
Machine Intelligence Garage is a unique programme that supports machine learning/artificial intelligence ventures across the key areas of need. As an early-stage startup, above and beyond getting access to computation power it wants to benefit from the guidance/support of thought-leaders and experts on technical and ethical challenges of building an ML/AI-first start-up. Lastly, it anticipates to have significant networking and collaboration opportunities with other cohort members, which will amplify its learnings and supercharge its progress.The deliberations and discussions Premind AI had during the application process has helped it revisit some of its technical assumptions, sharpen its thinking on its business proposition and consider broader ethical implications of its proposition.

Raven is an intelligent crawler developed to provide an automated search mechanism for videos with extremist contents on social platforms. Raven offers to increase throughput for internet companies to locate multimedia content on their social platforms, reducing time and stress for human analysts, and provide links and video analytics to law enforcement and counterterrorism.
Raven uses an ensemble of machine learning algorithms to analyse, index and extract relevant objects from videos. These machine learning algorithms need to be constantly updated and tuned based on user feedback. The Machine Intelligence Garage will provide us with the perfect tools to develop and tune our machine learning pipeline so we reduce classification time and incorporate user feedback in real-time.

The Machine Intelligence Garage programme has been extremely useful for Raven to understand all the possible ramifications of using machine learning algorithms in the real world. The programme has helped it to identify issues in its data collection and data labelling methods. It has also been extremely helpful on identifying some of the ethical issues and considerations it will need to take into account when developing Raven further.

WildMeta is using machine learning to build smart bots that work with and for humans to help them develop their skills at competitive video games. With bigger and bigger cash prizes, esports is continuously growing, which put pressure on players who increasingly rely on professional coaches and online analysis services to guide their progress. It believes that more could be done to support players through a new proactive esports training solution that provides the right virtual opponents to improve, creates a safe space to experiment with new strategies, and helps overcome their weaknesses.

WildMeta applied to the Machine Intelligence Garage programme to benefit from Digital Catapult’s expertise, network and resources and take part in discussions and workshops around AI as well as meet fellow founders in the space.

Yepic help brands build deeper relationships with their customers, staff and partners by making personalised video creation easy and affordable with the power of AI. Yepic is developing an AI-powered video generator that allows brands to create videos from only an audio or text file. It is making mass video production as easy as sending an email. It only needs to analyse a 3 min video clip, to then generate a digital version of the actor that can be algorithmically controlled using a computer program. Yepic’s Technology is able to create lip movements and emotions from a sound without any actor being on site. The emotions and head motion are learnt from the training footage – think Mailmerge for video Yepic’s technology allows you to turn first-party data like name, in-app progress or loyalty into a personalised video on the fly. The Machine Intelligence Garage robust ethics framework has been an insightful resource when planning new features for Yepic. In addition to this without the generous support of AWS, GCP and Nvidia its R&D would grind to a halt.

Cloud Infinity Projects Ltd trading as Advanced Infrastructure is developing frictionless billing platform that allows customer assets to switch to the lowest cost energy supplier every 15 minutes. The platform is built on Energy Web Foundation software development kits with with machine learning algorithms enabling assets to autonomously and frictionless switch between suppliers.

How has Machine Intelligence Garage helped the company?
The company applied to the programme in order to accelerate its development cycle and demonstrate traction to seed investors. Referencing selection onto the programme has helped it demonstrate traction and has led to an offer of investment and an accelerator program. Conversations around ethics and compute choice have made us take decisions faster.

Aistetic is building a B2B tech platform for retailers to make tailoring accessible to more people, leveraging computer vision, deep learning & advanced visualisation. It wants to reduce online returns and increase conversion for retailers. Its purpose is to do this sustainably, reducing the environmental impact of online returns ensuring that shoppers buy clothes that fit them.

How has Machine Intelligence Garage helped the company?
Aistetic applied to Machine Intelligence Garage as it requires computation power and cloud solutions for the commercialisation of its computer vision and deep learning technology.

Atlas AI is making personal training more accessible by automating technique feedback in the gym. Users can get instant video analysis through IoT cameras and suggestions for improvements. Gyms can leverage this to help drive retention, increase new members and upsell personal training. Above all, Atlas AI aims to get more people to prioritise technique when they exercise.

How has Machine Intelligence Garage helped the company?
Atlas AI applied to the Machine Intelligence Garage in order to meet fellow IoT and machine learning startups. This network is invaluable, as learning about its respective experiences can help everyone in the community grow faster. Cloud computing credit will allow the company to accelerate the rate at which it develops new features, which will help it adapt to address changing customer needs. Finally, Machine Intelligence Garage will connect it to a network of investors in the space who will be able to give it guidance on how to scale Atlas AI.

reachAI’s mission is to democratise video analytics through Computer Vision with AI/deep learning and to bring state-of-the-art analytics to the wider market, with initial focus on sports.

How has Machine Intelligence Garage helped the company?

Machine Intelligence Garage was able to support reachAI in gaining exposure to state-of-the-art tools, hardware and libraries for AI/deep learning. It benefitted from cost savings through free access to powerful GPU hardware and time savings as deep learning training loads were executed much faster on top of the line GPU hardware.

Elemendar has built an AI analyst for Cyber Threat Intelligence (CTI), the information that tells cyber defenders about the threats they defend against. Elemendar makes a CTI report written for humans usable for defence in a minute rather than an hour. Graduate of the first GCHQ/NCSC Cyber Accelerator, current member of LORCA Cohort 3.

How has Machine Intelligence Garage helped the company?

Elemendar applied to MI Garage in order to experiment with and train more models using different techniques in parallel and more quickly, as well as to allow the company to iterate through hyperparameter tuning to build more accurate, more intelligent models sooner than we otherwise would.

Through Machine Intelligence Garage Elemendar accessed AI ethicists whose framework for answering ethical challenges helped it understand the risk of its tech being used to help drive attacks rather than to defend, and as a potential conduit for attack if we’re being used to spread disinformation.

DIEM Consulting was founded in 2002 as a niche strategy consultancy and in 2011 spun-out DIEM Analytics to focus on decision aids and creating augmented intelligence for complex situations. DIEM Analytics’ clients were increasingly asking it to be trained on the use of combined stakeholder engagement with modelling and simulation approach. By 2017, DIEM Analytics had identified a niche in the market for developing Artificial Intelligence-based decision tools. DIEM Analytics enables better decision making by combining visual representations of a problem and small-group interviews to capture users’ experiences in different scenarios, creating ‘explainable AI’, augmented intelligence and ‘sparse data’ AI. Clients include the MoD, research organisations and leading defence companies.

Digital Taxonomy provide software that combines the power of AI and human interaction to help automate and streamline the market research industry

Photogram.ai, is a computational photography startup specialising in AI-accelerated hardware and software for modern photography. It is building an entirely new device; an AI-accelerated computational camera, giving creatives necessary, open and modern tools to enhance their art form. Its project crosses the boundaries between deep tech and the creative industries with a mission to empower creatives with technology. Photogram.ai is at the intersection of computer vision, artificial intelligence, IoT, as well as creativity, photography and consumer goods. It was founded on the Entrepreneur First June 2019 cohort and has now incubated at UCL Innovation & Enterprise’s Hatchery and is part of the NVIDIA Inception Program for AI start-ups revolutionising industries.

Photogram.ai, is a computational photography startup specialising in AI-accelerated hardware and software for modern photography. It is building an entirely new device; an AI-accelerated computational camera, giving creatives necessary, open and modern tools to enhance their art form. Its project crosses the boundaries between deep tech and the creative industries with a mission to empower creatives with technology. Photogram.ai is at the intersection of computer vision, artificial intelligence, IoT, as well as creativity, photography and consumer goods. It was founded on the Entrepreneur First June 2019 cohort and has now incubated at UCL Innovation & Enterprise’s Hatchery and is part of the NVIDIA Inception Program for AI start-ups revolutionising industries.