When it comes to AI (Artificial Intelligence), VCs (venture capitalists) continue to be aggressive with their fundings. During the third quarter, 965 AI-related companies in the US raised a total of $13.5 billion. In fact, this year should see a record in total fundings (last yearâs total came to $16.8 billion).
Some of the deals have been, well, staggering. Just look at the $1 billion that Microsoft shelled out for an equity stake in OpenAI (the company is one of the few that is pursuing Strong AI).
So what has been the result of all this activity? What have been the breakthroughs for AI this year?
There are plenty:
Guy Caspi, the CEO of Deep Instinct:
âOne area that is particularly interesting is Generative Adversarial Networks (GAN). In May 2019, researchers at Samsung demonstrated a GAN-based system that produced videos of a person speaking with only a single photo of that person provided. Then, in August of this year, a large dataset consisting of 12,197 MIDI songs each with their own lyrics and melodies were created through neural melody generation from lyrics by using conditional GAN-LSTM.â
Sheldon Fernandez, who is the CEO of DarwinAI:
âThough somewhat unheralded in mainstream circles, the rise of transformer based large scale language models such as BERT and ROBERTa represented a hugely significant development in the realm of Natural Language Processing (NLP), achieving state-of-the-art performance in General Language Understanding Evaluation (GLUE). Considered by many to be the âImageNet momentâ for NLP, numerous teams, including ours, have built on such work for critical tasks as Fake News Detection.â
Seth Siegel, who is the partner of Artificial Intelligence & Automation at Infosys Consulting:
âTensorflow when released took the data science world by storm, however, like all great technologies the growing pains of having to integrate multiple independent components like Keras provided poor developer experiences and slowed down projects. With the release of Tensorflow 2.0 and the integration of Keras, data scientists and developers are rapidly developing models and eliminating bottlenecks in the model build out process. We are excited to see the rapid iteration of the underlying tools and environments that help build the AI world.â
Krishna Gade, who is the CEO of Fiddler Labs:
âAI Explainability has hit mainstream. Techniques based on classical economics such as the Shapley value have gained prominence and companies are adopting them to validate AI models, explain the performance to business stakeholders and also to understand the worth of the training datasets supplied to the AI models.
âReinforcement Learning is also gaining momentum, especially in places where there is a limited supply of training data. Weâve seen a pretty cool demonstration of a Rubikâs cube built by a Robotic hand using RL.â
Chad Meley, VP of Marketing at Teradata:
âRetail vision had a break-out year in 2019, marking the beginning of âbrick and mortarâ stores becoming an essential part of enterprise digital transformation. Retail innovators are exploring AI-enabled computer vision to extract meaningful insights from CCTVs such as store traffic patterns and dwell times. Unlike many enterprise AI use cases that improve the effectiveness of previous generation models, retail vision has a certain coolness factor resulting from generating insights out of previously intractable data sets. The only way to observe people moving, or how long it took for an employee to help a customer was to have another person physically observe the situation. Now we can understand whatâs happening in the physical environment in a way that scales and operates continuously.
âSome of the use cases include sentiment analysis, which uses facial recognition to sense when a customer is frustrated or delighted, and associate time motion, where AI tracks and analyzes sales associatesâ activitiesâsuch as how much time they spend engaged with customers, how much time they spend on the store floor vs the back room, how much time they generally spend performing value added activities and so on.â
Flavio Villanustre, the VP of Technology & CISSO atLexisNexis Risk Solutions, which is a part of RELX:
â2019 has been an exciting year for AI and Deep Learning. There have been a number of great innovations that are setting the path for breakthroughs in the upcoming future, such as Geoffrey Hintonâs Capsule Networks, which provide computers with the ability to identify visual patterns far more accurately that the more traditional convolutional network filters that were being used. By the way, in case you didnât know already, Hinton was one of the three top world researchers honored with the 2018 Turing Award for advances in deep neural networks. The Capsule Networks can be directly applied to critical applications of AI, such as medical diagnostics, structural analysis, and autonomous and assisted driving vehicles. Additionally, these and other advances in AI are helping bring new capabilities to Computer Aided Design (CAD), particularly in very complex and tedious areas such as nanoparticles, molecular design and materials research. A different type of innovation came from the commercial availability of Quantum Computers, with even an offering from Amazon Web Services called BRAKET meant to make Quantum Computing as a service widely available. Quantum Computers can provide such a massive speedup of certain AI related tasks, that they could be a game changer when it comes to using AI applied to general problems at speeds that canât be achieved with more traditional systems.â
David Benigson, CEO of Signal AI:
âThis year we have seen artificial intelligence evolve into augmented intelligence; a technology specifically created to assist enterprise decision-making rather than to replace it. Augmented intelligence drives business results by monitoring and summarizing to create intelligent insights that help surface previously unseen issues.â
Florian Douetteau, the CEO of Dataiku:
âMachine language is getting democratized: this year weâve seen more and more people from the business empowering themselves and collaborating with technologies to build their own AI solutions.â
Tom (@ttaulli) is the author of the book, Artificial Intelligence Basics: A Non-Technical Introduction.
This is a syndicated post. Read the original post at Source link .