Artificial intelligence’s (AI) influence is engulfing large corporations with data and, more specifically, predictive inferences and automations are going to be the wave. Static analytics have hit an inflection point, whereby core AI technologies are providing better accuracy, latency, concurrency and stability to everyday processes, systems and computations.
Moore’s law has enabled the better access to sensors, graphical processing units (GPUs), data and talent. As a result, the following areas in AI have birthed access to a suite of new services which have the prospect of having an economic boost as high as 26% to overall global GDP by 2030, valued at approximately USD 15.7 Trillion.
Top 4 Most Popular Ai Articles:
1. Deep learning for sensor-based human activity recognition
2. How to build a deep learning server based on Docker
3. Transfer Learning: retraining Inception V3 for custom image classification
4. How ethical is Artificial Intelligence?
- AI Clouds: Lego blocking cloud based services with developer kits, large general purpose AI companies are enabling developers to deploy algorithms via SDKs within their cloud hosted platforms. From Microsoft Azure AI platform all the way to Amazon’s AWS AI Offerings, these organizations provide pre-trained models, GPUs and storage that are necessary for more effective continuous deployment, testing and quality assurance (QA).
- AI Languages: Beyond software applications to onboard users onto AI platforms, companies are standardizing new languages to familiarize developers to continually build using their libraries. Uber’s AI Labs, for example, released their own probabilistic python offshoot programming language, Pyro. Wit.ai is another language for developers to build cross device applications.
- AI Chipsets: Huawei, Apple, NVIDIA and Qualcomm are focused on building ‘SoCs’ or Systems on Chips. These system architectures come with better processing speeds and security protocols around data. Alphabet has launched its own GPU called the Tensor Processing Unit (TPU). TPU uses the Tensor Flow framework for a high volume of low precision computation (e.g. as little as 8-bit precision) with higher IOPS per watt, and lacks hardware for rasterisation/texture mapping.
- AI Bots: In more recent months, bots have become commonplace plugins to automate customer inquiries. From automating back offices to booking hospitality, bots have had a huge impact in streamlining enterprise operations and headcount. X.ai and Clara Labs are scheduling bots that handle consumer and enterprise meeting requests. Far from complete automation with the systemized use of humans in the loop (HITL), bots are prospectively among the initial use cases to be completely automated.
- AI Marketplaces: Sometimes misconstrued as predictive e-commerce platforms, AI marketplaces are markets for AI algorithms. Think of it as a marketplace for GitHub code. Market players dominating this space include In-Q-Tel funded and Seattle based Algorithmia, where developers can put up their algorithms for sale, PrecisionHawk, where drone operators can purchase precision agriculture related algorithms and DataXu, where marketing professionals can purchase proprietary content algorithms.
- AI APIs: With AI deployments becoming more everyday, companies incorporate smarter capabilities in their systems via APIs in machine learning, natural language processing and computer vision. Dialogflow provides text based conversational interfaces to connect to Google Assistant, Amazon Alexa and Facebook Messenger. Clarifai and Rasa are noteworthy companies with APIs designed for computer vision and natural language understanding applications respectively.
- AI Digital Assistants: Digital assistants from GAMA (Google Assistant, Amazon Alexa, Microsoft Cortana and Apple Siri) are ubiquitously used across devices to assist in all personal inquiries. Digital assistant adoption wars will be a GAMA conquest, won on the basis of support for OEM manufacturers like speakers, television and smartphone manufacturers, integrations with digital services like Netflix, Hulu and Spotify and access to data via downloaded applications.
- Enterprise Automated Machine Learning (AML): AML driven enterprise modeling is utilized in exploratory data analysis, feature transformations, algorithm selection, hyper-parameter tuning, model diagnostics and data leakage solves. It tends to be most useful for regression and classification problems involving tabular datasets that can increase a data scientist’s productivity, often by an order of magnitude. Tools include TPOT, Auto-Sklearn, Auto-Weka, Machine-JS and DataRobot.
- Enterprise Automated Natural Language Processing (ANLP): ANLP can be divided into automated natural language understanding (ANLU) and automated natural language generation (ANLG). NLU allows researchers to quantify and learn from all of that text by extracting concepts, mapping relationships and analyzing emotion. NLG allows researchers to build narratives out of big data sets that are customer/ consumer ready. NLG may be viewed as the opposite of NLU. Parlo’s advanced enterprise NLU Broca (Used by New Balance and Salesforce) incorporates semantic parsing, grammar engineering, and machine learning capabilities under one offering. Arria (Used by Accenture, Tableau and Deloitte) and Narrative Science (Used by PwC, Groupon and Deloitte) provide enterprise level NLG solutions that solve for global contextual and narrative issues affecting end level clients.
- Enterprise Automated Computer Vision (ACV): ACV can be broken into visual search (searching on indexed data), visual discovery and visual forecasting. While visual search and discovery focus on direct recommendations based on vector patterns on indexed or uploaded data (Example: Visenze), visual forecasting incorporates behavior analysis on image vector patterns to provide insights on actions of actors/ items that exist in-image and in-video. To better understand this, for example, looking potentially at geospatial information systems (GIS) imagery across multiple historical data points, once can predict the origination of environmental or climate breakouts (Example: Planet Labs).
- Industrial Robots: AI is powering the next generation of robots to assist in the automation of manual tasks across precision agriculture (Example: Ecorobotix), marine exploration (Example: OpenROV), construction (Example: Construction Robotics), inventory management (Example: RobotIQ) and telemedicine (Example: Da Vinci by Intuitive Surgical). Most companies are currently fixing the headwinds caused by the computation-movement gap, known as Moravec’s Paradox, to build out other applications. Vicarious is focused on solving part of this paradox, building the all encompassing hardware OS which tackles combinatorial data, task and cognitive computations.
- Aerial Passenger Vehicles: Autonomous flight is the new paradigm being tackled by the brightest minds to reduce the randomization of piloting quality and optimize flight mobility via algorithmic consistency. Boeing has started rethinking its complete supply chain and design roadmap to incorporate AI. Workhorse, Larry page funded KittyHawk and Intel’s Volocopter are among a suite of early movers lobbying for better Federal Aviation Authority guidelines. With the incorporation of EV technology by the likes of Rolls Royce, Airbus and Siemens, these new battery powered air taxis and hybrid planes are poised to change aviation.
- Aerial Freight Vehicles: Unmanned aerial vehicles (UAVs) in the delivery of small to medium payload goods is older news. Drone company Matternet was an early mover in delivering medical supplies and postal services in Switzerland. Since then, mainstream players like Amazon and UPS have been operating independently or partnering with companies like Workhorse to deliver goods with specific payloads like under 10 pounds for flights lasting up to 30 minutes. Other large scale players like FedEx, to up its 22% market share in logistics, are utilizing land based autonomous trucks via Peloton to test drive self-driving tractor trailer platoons.
- Autonomous Vehicles: Google’s Waymo, Uber and Lyft have been test driving autonomous vehicles for awhile in the streets of San Francisco. As California leapfrogged pan national legislation to enable statewide testing of autonomous vehicles, adoption will continue to evolve as algorithmic efficiency bypasses safety concerns. Companies like Renovo (partnered with Samsung), Blackberry’s QNX (partnered with Baidu) and NVIDIA (NVIDIA Drive) are focused on building self driving platforms and OSs to allow OEM auto manufacturers to focus on integration.