R and D
Case Studies
AI system predicts risk of diabetes with an 88% accuracy rate in tests
Shanghai’s Ruijin Hospital has partnered with the Chinese AI startup 4Paradigm to apply AI in healthcare, particularly chronic health conditions. They have tested an AI-backed diabetes prediction and management product, which they hope will help them identify patients at risk of developing diabetes up to 15 years in advance. The system showed an 88% accuracy in tests on information from 170,000 people.
Linked use case: Predict risk of condition developing at an early stage
ARUP saves 790 engineering hours using machine learning to detect utility clash points planning a light rail system for Auckland
Arup, in a joint venture with Jacobs, was selected to plan a new light rail system for the city of Auckland. The assessment of utility systems clashing at different locations along the proposed rail line was automated using supervised learning algorithms, reducing the amount of engineering time which would have been required for manual checks by 790 hours.
Linked use case: Collate and evaluate site data for architectural planning
Action Against AMD and Benevolent AI aim to find treatments for age-related macular degeneration (AMD) that causes blindness using machine learning
Benevolent AI and a group of four charities, Blind Veterans UK, Fight for Sight, the Macular Society and Scottish War Blinded, have partnered to find treatments and a potential cure for AMD. Benevolent AI''s machine learning technology will be leveraged to analyse existing scientific papers, clinical trials information, images, formulas, patents and any other knowledge we have on age-related macular degeneration (AMD) to uncover potential patterns, connections and point researchers towards important research areas.
Linked use case: Accelerate drug discovery by automating research stages and data integration and analysis
Adidas learns from its consumers'' design creation to better anticipate future demand trends using machine learning
With the use of machine learning, Adidas is able to reduce the typical 18 month timeframe of turning trends into commercially saleable shoes to just 24 hours. It does so by letting consumers design their own shoes in its prototype SpeedFactory, from which then the product ships immediately. As a result, by analysing the co-created designs with machine learning, the company gains insight into future trends and can efficiently anticipate future demand.
Linked use case: Enhance product and service offering
Adobe Research team investigates current approaches in machine learning for automating data cleansing and finds them inadequatee
Researchers from Adobe Research investigate the ability of metric learning techniques to automatically clean data. These approaches assume that datasets can be described in pre-defined ways, and that cleaning methods will work similarly well for similarly defined datasets. Ultimately, however, they find this not to be the case with the implication being that current cleaning standards are inadequate.
Linked use case: Automate data cleansing and validation
Adobe is developing technology to detect manipulation in images using a deep neural network
Adobe revealed that it is currently working on developing an AI system that spots whether an image has been altered artificially. It does so using a deep neural network which is trained on existing data sets of altered images.
Linked use case: Detect manipulated or falsified media
Adore Me generates product insights by determining customer sentiment with a 92% accuracy based on natural language processing analysis of thousands of reviews
Adore Me, an ecommerce lingerie retailer, analyses 1000s of customer reviews and feedback through natural language processing with a 92% classification accuracy. This helps them understand popularity of products and opinions resulting in product insights and improvements.
Linked use case: Enhance product and service offering
Aidyia fully automates hedge fund trading using genetic evolutionary and other algorithms claiming a 2% return
Aidiya launched a fully automated hedge fund in 2015 that trades in the US stock market based on evolutionary and other algorithms. Their aim is to predict stock price movements over the longer term using data from various sources. In two years the fund has achieved 2% returns which is less than many human managed funds.
Linked use case: Automate through a robo-advisor client portfolio investments and rebalancing recommendations based on defined investment strategies
Aipoly app identifies objects and uses voice to describe them to visually impaired people
Aipoly has launched an app by the same name aimed to assist visually impaired people understand their surroundings. Users point their phone camera towards an object and without taking a picture of it, the app can identify and label objects which it then reads to them. It does not require an internet connection and can also be of use to people learning a new language.
Linked use case: Assist visually impaired people by describing images or reading text
Airbnb increases booking conversions by 5% using machine learning to match property hosts'' guest preferences with consumers
AirBnB wanted to better better understand what accommodation requests are accepted or rejected by property hosts. Using logistic regression techniques they modelled whether hosts preferred stays that resulted in limited gap nights, lots or fewer stays, and also how far in advance the bookings were made. AirBnB changed their consumer search results to emphasise host properties that would more likely be accepted. This led to a 5% increases in booking conversions. This demonstrated a two-sided marketplace with search results that were a function of not only consumer but host preferences.
Linked use case: Match expectations from both sides of a 2-sided online market
Airbus used deep machine vision to detect clouds in satellites decreasing the error rate from 11% to 3%
Airbus used convolutional neural networks to detect clouds in satellite imagery. They decreased the error rate from 11% to 3%, a 72% improvement. They used the TensorFlow ML framework and results were obtained in about a month.
Linked use case: Identify physical properties of scanned images
Android is planning to extend battery life in its mobile phones by implementing convolutional neural networks developed by DeepMind
Android is planning to implement research from DeepMind which analyses battery usage by apps on Android-powered phones. These convolutional neural networks predict consumption and adjust battery power accordingly, resulting in longer-lasting batteries.
Linked use case: Enhance product and service offering
Anglia Ruskin University researchers develop mobile system which detects tuberculosis with 98.4% accuracy
Researchers from Anglia Ruskin University test different machine learning methods to classify digital images created using biosensors for the presence of tuberculosis antibodies. The goal was to create a system which could process and classify the images on a portable phone. They were able to achieve a 98.4% accuracy doing so using a random forest method.
Linked use case: Diagnose known diseases from scans, images, biopsies, audio and other data
Anthem aims to predict the occurrence of allergies using machine learning
Anthem has partnered with Doc.ai to execute a 12-month trial to analyse how allergies affect people. The system is based on AI and machine learning algorithms to identify predictive models and identify patterns and on blockchain to ensure the privacy and anonymity of medical data.
Linked use case: Predict personalised health outcomes to recommend individual treatment approach
Apple iOS 10 creates thematic movies from user photos using facial recognition while remaining low latency
Apple has introduced a new user media management application for devices running its iOS 10 system. The app, called "Memories", uses facial recognition and other AI to automatically create thematic mini movies with soundtracks, all processed on the user''s device. This is achieved while keeping computational resource requirements low so as to avoid user inconvenience.
Linked use case: Create video and image based content optimised for customers and consumers
Aravind Eye Hospital identifies eye complications arising from diabetes with a 97.5% accuracy using machine learning
Dr Ramasamy Kim and his team at the Aravind Eye Hospital in Madurai have developed a system capable of identifying eye complications arising from diabetes. Over the past five years, the doctors have been training the system by examining 15,000 retinal images and indicating spots or other abnormal features. Based on that the system has achieved a 97.5% accuracy, in testing, in identifying such complications. The hospital has been in a pilot project to check the system against human assessment since 2016.
Linked use case: Diagnose known diseases from scans, images, biopsies, audio and other data
AstraZeneca plans to crack down on online sale of counterfeit drugs in China using machine learning and natural language processing
AstraZeneca teams up with Tencent and Alibaba to strengthen the fight against counterfeit efforts. Alibaba will put traceability codes on drug packages and through their app patients will receive personalised recommendations and healthcare services. Tencent will work with AstraZeneca to identify online counterfeit drug sale using natural language processing to identify suspect language.
Linked use case: Identify and manage potentially fraudulent activity
Autodesk redesigned iconic Scandinavian design chair achieving 18% less volume; 90% decrease in max displacement; and 79% in von Mises stress with generative design model
Autodesk redesigned the iconic Danish design chair by Hans Wegner to optimise its cost, development time, material consumption, and product weight. A generative design software, Dreamcatcher, was fed with the original design and produced thousands of variations which it then tested against the desired criteria. The final product benefits from 18% less volume; 90% decrease in max displacement; and 79% in von Mises stress, while it maintains its aesthetics.
Linked use case: Create new products
Autonomous Healthcare detects different types of ventilator asynchrony in ICU patients with machine learning
Autonomous Healthcare has developed technology that can manage a patient''s ventilation in the ICU. Using machine learning, it is able to detect anomalies between the patient''s own and a mechanical ventilator''s inhalation and exhalation patterns. The system was trained on data of waveforms from patients on ventilators and learned the signatures of different asynchrony types. In its first assessment on data it achieved the same accuracy as human doctors, and it is now being tested with real patients.
Linked use case: Alerting and diagnostics from real time patient data
Aviva offers up to 30% discounts on insurance premiums based on actual driving behaviour that is monitored and analysed through machine learning
Aviva offers up to 30% discounts on insurance premiums based on machine learning analysis of actual driving behaviour. With an app Aviva is able to measure individual customer driving acceleration, cornering and braking. Driving behaviour is correlated to risk and premiums through advanced analytics. Those with lower risk driving are offered lower premiums.
Linked use case: Enhance motor insurance by predicting driving behaviour through the use of language in Facebook posts
BBC is developing interactive fictional programs utilising the natural language generation capabilities of smart home assistants such as Alexa and Google Home
BBC plans to develop a natural language processing engine to offer interactive programmes. The initiative is titled ''Talking with Machines'' and will explore content, interaction and software development patterns. The first programme in this series, an interactive audio drama has been already broadcast in November 2017. The programme is now being expanded.
Linked use case: Automate generation of conversational media content
BERG is attempting to identify genetic predisposition to certain conditions using machine learning with promising Phase I study results
BERG has developed a platform to swiftly analyze patient biology and identify biomarkers using machine learning. Their platform was used in Phase I study to discover molecular markers identifying patients more likely to benefit from the medicine, thus applying a precision medicine approach.
Linked use case: Identify and validate a molecule to target with a drug compound
BERG is developing targeted cancer drugs using machine learning with promising Phase I study results
BERG has developed a platform to swiftly analyze patient biology and identify biomarkers using machine learning. Their platform was used in Phase I study to identify patients more likely to benefit from the medicine, applying a precision medicine approach. They have used the platform to develop a new cancer drug and reduce chemotherapy-induced alopecia.
Linked use case: Identify and validate a molecule to target with a drug compound
BMW will self-diagnose faults and issues limiting car performance and collect data from connected vehicles with machine learning analytics
BMW is planning to use IBM''s cloud infrastructure to deploy its new platform for collecting data from connected cars. IBM announced it''s become a "pilot partner" with the German automaker.
Linked use case: Discover anomalies across fleet of vehicle sensor data to identify potential risks
Babylon Health claims 82% accuracy for video medical diagnosis based on machine learning and natural language processing
Babylon health has developed an inference engine based on machine learning to diagnose disease. It refers to the knowledge graph built from medical knowledge. The patient interface is an app with video capabilities which uses natural language processing. Babylon has a partnership with the NHS in the UK and is being piloted in Rwanda. Performance and data privacy concerns have been raised.
Linked use case: Provide first line of medical advice online through chatbot