For many years, AI has eluded smartphones owing to its own resource-heavy nature. With the introduction of tools like TFLite and other frameworks, machine learning models became lighter; light enough to be boarded onto the limited space on a mobile or an edge device. This allowed app developers to incorporate models into their apps for better experience. From Maps to face filters, from e-learning to contact tracing during COVID-19, AI came in quite handy. In this article we list few of the most popular AI based apps (not necessarily created this year) that have made their presence felt this year.
COVID-19 Sounds App
One of those apps that were targeted at keeping the pandemic under check. Launched by the University of Cambridge, the app allows users to upload short recordings of cough and breathing and report symptoms to help researchers from the University of Cambridge detect if a person is suffering from COVID-19. The app will collect some voice samples, while you read text on the screen and a few seconds of breathing and coughing. According to the website, the aim is to collect data to inform the diagnosis of COVID-19 by developing machine learning algorithms, based primarily on sounds of their voice, their breathing and coughing.
The app will collect some basic demographics and medical history data, as well as some voice samples (while you read text on the screen) through a questionnaire and a few seconds of breathing and coughing through the phone microphone and will additionally collect one location sample. However, the creators also assured that their app wouldn’t be tracking users.
Microsoft Math Solver
Microsoft’s Math Solver app is designed to solve various math problems, including quadratic equations, calculus, and statistics. Users just have to scan the problem on their math notebook, and Microsoft’s app powered by computer vision algorithms finds the solution. The app not only solves the formula, but it also shows you how you can achieve the same result using different methods. It also provides additional learning resources on topics, including videos.
Microsoft Math Solver supports 22 languages, including English, German, Spanish, Simplified Chinese, and Hindi.
AI Dungeon is an AI-based text adventure game that is built using OpenAI’s recent hit GPT-3 model. AI Dungeon leverages GPT-3 to decide what happens in a story no matter how crazy your action is. According to the creators, this is how the Multiplayer (Alpha) AI Dungeon works:
- AI generates a result after each action of the players.
- Actions and stories are all in third person rather than the second person.
- The host of a multiplayer game can act as a dungeon master and supervise the AI, modifying and fixing responses it gives.
Know more here.
An image manipulation app FaceApp allows its users to alter and transform their images using filters. For this, the app works on image recognition technology, that is key for facial recognition systems, and utilises deep learning for recognising the key features, like eyelids, cheekbones, jawline, nose bridge etc. of the human face to create those transformations.
Microsoft’s Seeing AI is built on AI-powered image recognition, and description technology, especially for those with visual impairment. This app assists the users to recognise all the visual elements in their surroundings through audio narration. The elements are identified through the camera, and an audio description is generated. The app automatically recognises contacts in the user’s phone if there are photos of that person and can also estimate the age, sex, race and emotion of people it does not know yet.
Other features include short text reading, document scanning, barcode recognition, scene preview, colour recognition, light detection, and the ability to describe images across all social media apps.
Read Along by Google
Read Along by Google is an Android app for kids above five years of age. It is designed to help them learn to read by giving verbal and visual feedback as they read stories out loud.
The app uses Google’s state of the art speech recognition technology to help develop literacy skills. Read Along helps kids with the help of an in-app assistant ‘Diya’. As kids read out loud, Diya uses Google’s text-to-speech and speech recognition technology to detect student’s reading progress.
Read Along has been designed to personalise the experience by recommending the right difficulty level of stories and games based on their reading level performance. Read Along is currently available in nine languages including English, Spanish, Portuguese and Hindi, over 180 countries.
Microsoft launched Lobe, which allows users to train machine learning (ML) models without writing any code.
Lobe has everything one needs to bring machine learning ideas to life. Users just have to show some examples of what one wants the app to learn, and it automatically trains a custom machine learning model that can be shipped in your app. Currently, Lobe supports image classification and further, Microsoft plans to extend its capabilities to train other models.
Qanda.ai was launched by Seoul-based edtech startup Mathpresso. The app claims that users can scan math problems they don’t understand, and will get solutions in just five seconds. QANDA (stands for ‘Q and A’), allows students to receive instant answers and customised learning sources at zero or minimal cost.
The app recognises text and mathematical formulas in a photo using optical character recognition (OCR) technology and provides solutions to a student’s question with high accuracy. QANDA has over 5.7 million monthly active users in over 50 countries. The app currently offers seven languages – Korean, English, Spanish, Japanese, Vietnamese, Indonesian, and Thai.
Mubert: AI Music Streaming
Mubert is a streaming app that generates unique music personal to each listener. It leverages AI to skim through a huge database of human-made sounds in different moods and genres to create never-ending music and generate music customised to the user. The Mubert community that consists of thousands of creators call this genre, “Generative Music”.
Created by the students of IIT Bombay, AIRScanner is an AI-based Reading Assistant and PDF Document Scanner app, powered by its unique AI Dictionary and Narrator. AIR Scanner uses AI to understand words in a page and on just a single tap on word boxes, the user gets its full-fledged meaning, similar to what we have seen on Kindle app.
(Note: The following list is in no particular order and has been compiled based on the number of downloads in the app stores, popularity, reputation of the creators with regards to AI research etc., If you think we have missed any, please add them in the comments.)