AI technologies are some of the most complex and exciting topics going into the new year and so I’ve put together a list of the top 10 artificial intelligence trends to look out for in 2018.
THE UBIQUITOUS-TO-BE DEEP LEARNING THEORY
Deep neural networks, which mimic the human brain, have demonstrated their ability to “learn” from image, audio, and text data. The combination of neural networks (enabled by the cloud), machine learning technology, and massive data sets (the internet), has made Deep Learning one of the most exciting AI sub-fields recently. Yet there’s still a lot we don’t yet know about deep learning, including how neural networks learn or why they perform so well. Understanding precisely how deep learning works will enable its greater development and use. Following core deep learning trends will dominate in 2018:
- A new theory that applies the information bottleneck Principle to deep learning will help us know how neural networks learn. It suggests that after an initial fitting phase, a deep neural network will “forget” and compress noisy data—that is, data sets containing a lot of additional meaningless information—while still preserving information about what the data represents.
- Given big data sets, Deep Learning algorithms are great at Deep Pattern Recognition, and enable things like, speech recognition, image recognition, natural language processing. Example DeepFace, (Facebook).
- Convolution Neural Networks will be the prevalent bread-and-butter model for DL systems. RNNs and LSTMs with its recurrent configuration and embedded memory nodes are going to be used less simply because they would not be competitive to a CNN based solution.
Reference Links
https://sameerdhanrajani.wordpress.com/2017/10/14/design-thinking-behavioral-sciences-strategic-elements-to-building-a-successful-ai-enterprise/
https://sameerdhanrajani.wordpress.com/2017/06/03/drones-driverless-cars-ai-at-the-core/’
DESIGN THINKING DISRUPTING AI ITSELF
Design thinking is defined as human-centric design that builds upon the deep understanding of our users (e.g., their tendencies, propensities, inclinations) to generate ideas, build prototypes, share what you’ve made, embrace the art of failure (i.e., fail fast but learn faster) and eventually put your innovative solution out into the world. Although cognitive design thinking is in its early stages in many enterprises, the implications are evident. Eschewing versus embracing design thinking can mean the difference between failure and success.
The major trends in Design Thinking for 2018 will be:
- Design Thinking will see divergence from More Powerful Intelligence To More Creative Intelligence. While algorithms can automate many routine tasks, the narrow nature of data-driven AI implies that evolving into creative intelligence will enhance the human-AI augmented work philosophy.
- Cases of Ethical Quandries in Enterprise premise may evolve. Unintended algorithmic bias can lead to exclusionary and even discriminatory practices. Accordingly, across many fields, we can start thinking about how we create more inclusive code and employ inclusive coding practices.
- As part of CXO Strategy for Cognitive Design Thinking, CIOs will introduce them to their organizations by first determining how it can address problems that conventional technologies alone cannot solve. They benefit from working with business stakeholders to identify sources of value.
Reference Links
https://sameerdhanrajani.wordpress.com/2017/10/14/design-thinking-behavioral-sciences-strategic-elements-to-building-a-successful-ai-enterprise/
AUTOML: MODEL CREATION WITHOUT PROGRAMMING
Called “AutoML” for “auto-machine learning,” it allows one AI to become the architect of another, and direct its development without the need for input from a human engineer. Automated Machine Learning (AutoML) systems started becoming competitive with human machine learning experts. AutoML systems had started replacing human experts for standard machine learning analyses in 2017 and will continue the trend in 2018. The governing trends in 2018 will be:
- AutoML will take over ML model-building process. Once a data set is in a (relatively) clean format, the AutoML system will be able to design and optimize a machine learning pipeline faster than 99% of the humans out there.
- All the methods of AutoML are developed to Augment data scientists’ tasks, not to replace them. Such methods can free the data scientist from nasty, complicated tasks (like hyperparameter optimization) that can be solved better by machines. But analysing and drawing conclusions still has to be done by human experts
Reference Links
https://sameerdhanrajani.wordpress.com/2016/12/31/sameer-dhanrajani-2017-data-science-analytics-trends/
https://www.kdnuggets.com/2017/01/current-state-automated-machine-learning.html
RISE IN NARROW AI IMPLEMENTATION IN STARTUPS AND ENTERPRISES
AI will continue to actively expanding its footprint within the Narrow AI enterprise. Executives will try to more fully comprehend what AI is and how they can use it to better capitalize on business opportunities by gaining insights into their data and engaging with customers more productively, thereby honing a competitive edge. Investment in AI Startups will also continue to climb as last year. To date, the market contains 2,045 AI startups and more than 17,000 market followers, with more joining by the day. Some major trends in 2018 will be:
- Need of large volumes of labeled data to train the system can be addressed by Lean Data Learning Techniques like transferring a model trained for one task or domain to another. Also, Augmented Data Techniques like synthesizing new data through simulations or interpolations helps obtain more data, thereby augmenting existing data to improve learning.
- Auditability and ‘explainability’ of AI will go mainstream. While Explainable AI (XAI) work is still in its infancy, enterprise AI platforms like Infosys Nia have started including auditability and basic visualization tools to take steps towards a system that doesn’t behave like a black box.
- High AI Startup Investments will continue to be the trend, across use cases within the industries. Last year, VCs struck 658 deals with AI companies, nearly five times the number four years before.
Reference Links
https://sameerdhanrajani.wordpress.com/2017/12/18/man-machine-ai-the-future-is-here/
https://sameerdhanrajani.wordpress.com/2017/08/20/how-startups-can-leverage-ai-to-gain-competitive-advantage/
https://www.darpa.mil/program/explainable-artificial-intelligence
DEEP REINFORCEMENT LEARNING: EMPOWERING ROBOTS
DRL is type of neural network that learns by interacting with the environment through observations, actions, and rewards. Currently, deep learning is enabling reinforcement learning to scale to problems that were previously intractable, such as learning to play video games directly from pixels. Deep reinforcement learning is poised to revolutionize the field of AI and represents a step towards building autonomous systems with a higher level understanding of the visual world. Hottest trends in DRL for 2018 are:
- DRL is the most general purpose of all learning techniques, so it can be used in the most business applications. It requires less data than other techniques to train its models.
- Deep reinforcement learning (DRL) will be used heavily to learn gaming strategies, such as Atari and Go—including the famous AlphaGo program that beat a human champion.
- This will become the cornerstone for robotization with DRL being the foundation of making robots learn. Deep reinforcement learning algorithms are also applied to robotics, allowing control policies for robots to be learned directly from camera inputs in the real world.
Reference Links
http://usblogs.pwc.com/emerging-technology/top-10-ai-tech-trends-for-2018/
CAPSULE NETWORKS: EMULATING THE BRAIN’S VISUAL PROCESSING STRENGTHS
Capsule networks, a new type of deep neural network, process visual information in much the same way as the brain, which means they can maintain hierarchical relationships. This is in stark contrast to convolutional neural networks, one of the most widely used neural networks, which fail to take into account important spatial hierarchies between simple and complex objects, resulting in misclassification and a high error rate. Expect to see the widespread use of capsule networks across many problem domains and deep neural network architectures. The Key trends of Capsule Networks in 2018 will be:
- Since Capsule Networks consider translation equivariance, they will become key in the development of Advanced Computer Vision.
- Capsule Networks are capable of learning by only using a fraction of the data that a CNN would use. But current implementations are much slower than other modern deep learning models. Coming year will witness capsule networks being trained quickly and efficiently.
Reference Links
https://medium.com/ai%C2%B3-theory-practice-business/understanding-hintons-capsule-networks-part-i-intuition-b4b559d1159b
https://kndrck.co/posts/capsule_networks_explained/
CONVERSATIONAL AI ENABLING NATURAL INTERACTION
As e-commerce (and retail in general) becomes increasingly global and competitive, business leaders understand that Conversational AI can be a valuable tool in reconnecting with consumers. Chatbots, in the form of virtual assistants and automated customer service reps, are becoming increasingly common across the industry. Meaningful applications of conversational AI are already quietly up and running, and as cost benefits continue to pile up, the trend will accelerate in 2018. Some of the key facets of Conversational AI in 2018 are:
- Enterprise Conversational AI will see mainstream adoption as 20% of firms will look to add voice enabled interfaces to their existing point-and-click dashboards and systems.
- To build good Customer Experience companies are turning to artificial intelligence. The new AI Chabot can help customers get the answers they need. However, instead of chatting with a human, customers are communicating with a machine that uses trends and previous knowledge to provide the right answers.
- (Conversational) AI is becoming the face of a brand. Some of the leading digital businesses are already securing significant advances in their use of AI for everyday dealings with the consumer. In only a few years, it’s likely that most interactions won’t require a keyboard. Instead, they will be based on voice, gesture and augmented or virtual-reality interactions.
Reference Links
https://sameerdhanrajani.wordpress.com/2016/11/04/sameer-dhanrajani-the-ai-powered-retail-revolution/
https://sameerdhanrajani.wordpress.com/2017/05/20/cx-ai-dca-how-ai-is-accentuating-customer-experience-leveraging-digital-chat-agents-chatbots/
ROBOTIZATION: AUTOMATION AND AUTONOMY
RPA uses traditional computing technology to drive its decisions and responses, but it does this on a scale large and fast enough to roughly mimic the human perspective. According to McKinsey analysts, 81% of the time that workers spend on manual labor can be passed on to robots; automation of data processing will save 69% of employees’ work hours, and automation of data collection. Technologies will change the requirements for employees: to supplement robots (rather than compete with them), people need to develop such qualities as creativity, emotional intelligence, and cognitive flexibility. Key 2018 trends are:
- Robotization will displace industries like supply chain, manufacturing, Healthcare & Transportation, aimed at making our lives easier. RPA is delivering more near-term impact, but the future may be shaped by more advanced applications of true AI
- Cloud base AI will learn from Big Data to enable human-like social robots that can perform usefully as personal assistants EXAMPLES: Kuka Robotics Boston Dynamics
- Robotization will further affect jobs, especially involving administrative tasks. World Economic Forum (WEF) states that by 2020 due to the integration of new technologies 7.1 million people will lose their jobs, mostly white-collar workers engaged in office and administrative routines.
Reference Links
https://www.solutionanalysts.com/the-most-influential-mobility-trends-to-look-out-for-in-2018/
INTELLIGENT THINGS FOSTERING EMBEDDED INTELLIGENCE
Intelligent things are physical things that go beyond the execution of rigid programming models and exploit AI to deliver advanced behaviors that interact more naturally with their surroundings and with people. AI is delivering enhanced capability to many existing things, such as IoT-connected consumer and industrial systems. This phenomenon is closely aligned with the emergence of conversational platforms, the expansion of the IoT and the trend toward digital twins. Major trends of Intelligent things in 2018 will be:
- AI will be embedded more often into everyday things, such as appliances, speakers and hospital equipment. In Smart Homes avenue, the launch of more affordable devices like Google Home Mini as well as Alexa will only serve to increase that comfort level.
- Embedded intelligence in industrial IoT and other business scenarios will rise. For example, today’s digital stethoscope can record and store heartbeat and respiratory sounds. Collecting a massive database of such data, relating the data to diagnostic and treatment information, and building an AI-powered doctor assistance app would enable doctors to receive diagnostic support in real time.
- Smartphones to become more like smart home devices, in terms of centralized voice usage instead of using individual apps. Apple’s delayed HomePod will drive smart home device adoption even further, and into affluent households.
Reference Links
https://sameerdhanrajani.wordpress.com/2017/03/24/the-best-practices-for-internet-of-things-analytics/
https://sameerdhanrajani.wordpress.com/2017/04/15/a-perspective-to-cios-on-emerging-cross-industry-digital-models-under-industrie-4-0/
DEEP BEHAVIOR: AI THAT UNDERSTANDS YOU
Much of the Behavioral Science aspect of AI till now has been more about observing the people’s behaviors and prompting meaningful actions based on the facts of triggers, engagement, and habits. This will further grow, and also open doors to concepts like perceptive behavioral responses, and influencing human behavior for social goals. Key trends for 2018 will be:
- AI in Behavioral Economics will increase in relevance to Applied AI. This will be used in the search for new “behavioral”-type variables that affect choice. This will also scale up research in tech-human interaction, which will require new knowledge from behavioral economics about attention and perceived fairness, and improve ethical decision-making.
- AI bots that nudge humans for behavior change, like altering human social behavior in groups will go mainstream. This can counter the flaw of local and personal focus that humans sometime exhibit which prevents the realization of solution to a social problem.
- Perception Intelligence, which can create almost real human like behavior states in virtual objects will rise, increasing its foothold in the Gaming , Entertainment and Animation Industry.
Reference Links
https://sameerdhanrajani.wordpress.com/2017/10/10/how-cxos-are-leveraging-ai-to-pivot-business-strategy-and-operational-models/
https://www.cs.toronto.edu/~duvenaud/courses/csc2541/slides/gan-applications.pdf
Demystifying Data Analytics, Decision Science & Digital 