Quantum Computing Algorithms for Artificial Intelligence

Dr. Amit Ray explains the quantum annealing, Quantum Monte Carlo Tree Search, Quantum algorithms for traveling salesman problems, and Quantum algorithms  for gradient descent problems in depth.

This tutorial is for the researchers, developers, students and the volunteers of the quantum computing team of the  Sri Amit Ray Compassionate AI Lab. Many of our researchers and students asked me to explain the quantum computing algorithms in a very simplistic term.  The purpose of this article is to explain that.

Earlier we have discussed Spin-orbit Coupling Qubits for Quantum Computing and the foundations of  Quantum computing and artificial intelligence.  This article is to explain the foundation quantum computing algorithms in depth in a simplistic way. Here we explained the concepts of quantum annealing, Quantum Monte Carlo Tree Search, quantum algorithms for traveling salesman problem and Quantum algorithms for gradient descent problems. 

Quantum Computing Basics

Rather than store information using bits represented by 0s or 1s as in classical digital computers do, quantum computers use quantum bits, or qubits, to encode information as 0s, 1s, or both at the same time. This superposition of states—along with the other quantum mechanical phenomena of entanglement and tunneling—enables quantum computers to manipulate enormous combinations of states at once. The measurement in a quantum computer is performed by measuring the states of the qubits (charge, spin, angular momentum etc.). 

Mainly there are four types implementation of qubits: spin qubits, photon qubits, flux qubits and charge qubits. We in Computational AI Lab, focus on implementing quantum computing with photonic qubits and solid state spin qubits in an integrated structure.

Key Equations of Quantum Computing

The foundation of quantum computing is the three basic theories of quantum mechanics:  the Non-Relativistic Time-Independent Schrodinger Equation, Relativistic many-fermion systems (the many-body Dirac equation), and Gauge field theories (quantum Yang-Mills theories).

The Schrodinger equation is the basis for non-relativistic wave equation used in one version of quantum computing to describe the behavior of a quantum particle (electron, photon or ion) in a field of force. There is the time dependent equation used for describing progressive waves, applicable to the motion of free particles. And the time independent form of this equation used for describing standing waves.

Components of Quantum Computing

Quantum computing has five key components: control and manipulation of the isolated quantum particles, measurement of the parameters of the quantum particles, error correction, quantum algorithms and scaling up for large scale implementation.  

There are three approaches to quantum computing: Quantum Annealing (QA),  Adiabatic quantum computation (AQC) and Gate-based quantum computing (GQC). Here, we will focus mostly on quantum annealing implementation of the algorithms. In the book Quantum Computing algorithms for Artificial Intelligence these approaches are discussed in details. 

Quantum Annealing: 

Quantum annealing is the set of meta-heuristic algorithms of quantum computing based on the concepts of quantum superposition, entanglement and tunneling. Quantum wave-functions and Quantum coherence are the key elements of quantum annealing.   The idea of Quantum Annealing is an offspring of the traditional thermal simulated annealing process, where the problem of minimizing a certain cost or energy function in a large space of configurations is solved, through a statistical mechanics analogy, by the introduction of an artificial temperature variable which is slowly reduced to zero in the course of a Monte Carlo simulation or Molecular Dynamics simulation. This device allows to explore the configuration space avoiding trapping into local minima, often providing a more effective and less biased search for the minimal “energy” than standard gradient-based minimization methods. 

Brain-Computer Interface and Compassionate AI

The purpose of Compassionate AI is to remove the pain from the society and help humanity. Artificial Intelligence with Brain-Computer Interface (BCI) or Brain Machine Interface (BMI) is a fast-growing emerging technology for removing pains from the society. Here,  Dr. Amit Ray explains how with the advancement of artificial intelligence and exploration of new mobile bio-monitoring  devices, earphones, neuroprosthetic, wireless  wearable sensors, it is possible to monitor  thoughts and activities of brain neurons  and serve humanity.

This research is going to be immensely  beneficial for the physically and mentally challenged people as well as for the people who are suffering from post-traumatic stress disorder (PTSD), and other mental disorders or brain problems. Over the last 5 years, technologies for non-invasive transmission of information from brains to computers have developed considerably.

Here, researchers focus to build a direct communication link between the human brain and the smartphones, earphone, computers or other devices. With BCI mind can speak silently with a smartphone or other devices.  Recent advancement of neuroprosthetic, linking the human nervous system to computers and providing unprecedented control of artificial limbs and restoring lost sensory function.

 BCI establishes two way communications between the brain and the machine.  One is  brain-computer interface and another is called computer-brain interfaces (CBI). BCI hopes to create new communication channels for disabled or elderly persons using their brain signals. 

Artificial Intelligence to Combat Antibiotic Resistant Bacteria

Antibiotic resistance bacteria is one of the key research area of our Compassionate AI Lab. Dr. Amit Ray explains how artificial intelligence can be used in combating these superbugs. Antibiotic resistance bacteria is becoming world’s biggest health crisis. We discussed here multi-agent deep reinforcement learning models for predicting behavior of bacteria and phages in multi-drug environments.  We call this model as DeepCombat. 

Antibiotic resistant bacteria are bacteria that are not controlled or killed by antibiotics. They are able to survive and even multiply in the presence of an antibiotic.  These bacteria currently kill an estimated 700,000 people globally each year – a death toll which could rise to 10 million a year by 2050 if we don’t act [1]. The main difficulty is that the bacteria are changing fast. They changing faster than we can change the drugs in response.

 Artificial intelligence is showing alternative means of fighting these deadly infections and killer bacteria. Multi-drug-resistant bacterial infections annually result in millions of hospital days, billions in healthcare costs, and, most importantly, thousands of lives lost. Artificial Intelligence for healthcare is progressing at an exponential rate.  We are evaluating here, the role of artificial intelligence in fighting these superbugs.  Especially, the use of AI for intelligent Phage therapy.

Grid Cells and Artificial intelligence for Assisting Navigation of Blind People

Designing automated navigation system for blind people is one of the key project  of our Compassionate AI Lab. Dr. Amit Ray explains how grid cell, place cell and path integration strategies with artificial intelligence can be used  for designing the navigation system for blind people. The project focuses on how image recognition, voice recognition and path navigation methods of artificial intelligence can be used  for automated navigation system for blind people. Artificial intelligence can be used for helping blind people in many ways. Here, we discuss about our recent project of using AI techniques for automatic navigation. 

Navigation for blind people in a complex environment is a challenging task. Visual Impairment makes the person depend on another person for all their works and daily chores. This project is aimed to help blind people and eliminate their pain.

The discovery of place cells and the grid cells in the hippocampus opened a new perspective for understanding  mammal navigation system. Applying grid cells, place cells and path integration concepts with artificial intelligence is the main aim of this project of developing automated navigation systems for blind people. 

AI Based Navigation System for Blind People

Our objective is to develop an automated navigation system for the blind people. Our long-term goal is to create a portable, self-contained system that will allow visually impaired individuals to travel through familiar and unfamiliar environments without the assistance of guide. Currently, the most widespread and used means by the visually impaired people are the white stick, helper or the use of guide dog; however both present some limitations. 

The goal of the project is to give blind persons the ability to move around in unfamiliar environments, whether indoor or outdoor, through a user friendly interface. The term blindness refers to the people who have no vision at all or people who have less vision. With the advancement in AI technology usage of image recognition, voice recognition and path navigation method it can be easier to send commands regarding directions to the blind people. 

Compassionate Artificial Intelligence Scopes and Challenges

With the advancement of AI and nuclear war technology, gradually mankind is moving towards a great threat. Compassionate artificial intelligence is the way to come out of that threat.  Here, Sri Amit Ray talks about how artificial intelligence, neural networks, deep learning, reinforcement learning and other machine learning technologies can be used for designing advance compassionate artificial intelligence systems. Dr. Ray discusses the scopes, issues and frameworks to include compassion, kindness and empathy in future AI systems.

Need for Compassionate Artificial Intelligence Systems

One may question why it is useful to study compassion in machines at all. Compassion is an important part of human intelligence. The main objective of AI is to serve humanity in an intelligent manner. As AI technology is improving, serving humanity on the surface level is not sufficient. AI can serve humanity in much better way in a much deeper sense.  Compassion, kindness and empathy are the components of higher human intelligence and to be true intelligent, artificial intelligence must incorporate them in the system. The scope and benefits of compassionate AI are many. Some of the requirements for compassionate AI are as follows:

Robots are leaving the realm of the industry and entering into our homes and workplaces. Humans not just interact with information but also with emotions and feelings. Robots are required to be soft and compassionate towards human.  

They are required for health care, education and removing the loneliness of people. 

Compassionate artificial intelligence systems are required to stop mass destruction weapon systems. The rapid decision-making capabilities of AI systems are now used for launching nuclear weapons in ships and submarines, especially for preventing counterattacks. AI is also used to control non-nuclear weapons including unmanned vehicles like drones and cyberweapons. In this competitive age, it is almost certain, that if some-country or someone turn AI into a full-fledged automated weapon system, everyone else will do that too. Initially, it may be just for their own defense but subsequently that will be used for mass destruction. The purpose of compassionate artificial intelligence systems is to develop inbuilt system that can stop and disobey the instructions of evil forces. Evil force can be a machine, algorithm or human.  

Classification of Compassionate AI systems:

In the book Dr. Ray argued that compassionate AI systems can be classified into three groups. They are; Narrow Compassionate AI, General Compassionate AI and Compassionate Superintelligence. Narrow Compassionate AI tries to remove pains at individual level. General Compassionate AI  tries to remove pains at social level, group level or country level.  Compassionate Superintelligence tries to remove pains at higher levels of humanity. 

Narrow Compassionate AI deals with narrow domains like looking after elderly people, helping blind people,  assisting healthcare services. General Compassionate AI  helps humanity at higher levels in an integrated infrastructure. It deals with social pains from higher levels. It will solve the social problems like monster government, political corruptions, terrorism, human exploitation,  unethical journalism of paid news and false news, social media based depressions etc.  Compassionate Superintelligence will save humanity from nuclear war, earth quake and other disasters. 

What is compassion?

We all feel compassion when we see our family or friends in distress, and even animals feel compassion when they see their offspring are in pain. Compassion is a form of emotional engagement that is beneficial to the sufferer and the humanity.  Compassion involves the  sharing of feelings of another as a means of coming to an understanding and appreciation for how they feel. Hence, compassion is the feelings, inner experiences and related efforts to remove the pain of others.  Compassion is an inbuilt human nature. It is the human potential to reach the higher evolution of consciousness. Compassion comprises of feelings, emotions, sentiment and understanding. 

Characteristics of Compassionate AI

Compassion involves emotional engagement with some agent. The agent can be a person, group, machine, country etc. It also involves understanding their world. AI is humanizing technology.

The Compassionate AI can care for someone by providing their needs but not necessarily moving into their personal world, walking in their shoes and sharing in their struggles, joys and challenges. Kanov et al. (2004) discuss that compassion consists of three facets: noticing, feeling, and responding. ‘Noticing’ includes being aware of a person’s suffering, either by cognitively recognizing this suffering or by experiencing an unconscious physical or affective reaction to it. ‘Feeling’ is defined as responding emotionally to that suffering and experiencing ‘empathic concern’ through accepting the person’s perspective and imagining or feeling their condition. Finally, ‘responding’ encompasses having a desire to act to alleviate the person’s suffering.

Implementing Compassionate Artificial Intelligence

Human emotions and feelings are like organic algorithms that respond to the environment. Now, machine learning algorithms are used to know human emotions by studying facial expression, words,  gestures, speech patterns, tone of voice or body language. Emotion API of Microsoft Azure is one such example.  IBM Watson Personality Insights is an example of human personality analysis. 

AI is advancing rapidly at emotional intelligence. Face-tracking software is already advanced enough to analyze the smallest details in our facial expressions.  The facial images are often analysed with deep learning algorithms which accurately classify them according to the feelings of the viewer. Emotion-Driven Reinforcement Learning is successfully for building narrow level compassionate models. 

Sources:

1. Compassionate Artificial Superintelligence AI 5.0, Amit Ray, 2018

Currently, deep learning modules of AI based systems lacks the emotional aspects of human intelligence. However, to fix the subjective issues like relationship, depression, anxiety and emotional issues future artificial intelligence based  systems like cyborgs require deep emotional intelligence modules.   AI is expanding and evolving itself in many technological fronts. It is not only limited by Deep learning algorithms, but expanding its horizons in deeper levels of human consciousness. Modern AI is tremendously successful for pattern recognition, voice recognition, face identification and machine learning. Self-driven cars are already on road in testing phase.  But, in today’s world AI is more needed in dealing with emotions like anger, impatience, disappointment, frustration, surprise, happiness, and gratitude. This article covers the scope, issues and challenges of AI for building emotional intelligence.

Need for Combining Artificial Intelligence with Emotional Intelligence

Human emotion is deeply associated with motivation, decision, evaluation, learning, character, intelligence, desires, and awareness. Thus, nearly all human psychological activities are subject to emotional influences and excitation. Self-awareness is the significance of the mental activities of a human being. Therefore, it should be one of the core ideas of AI with emotional intelligence.

Emotional intelligence is defined as the ability to recognize, understand and manage one’s own emotions and to recognize and influence the emotion of others. Obviously, emotional intelligence separates us from the machines. It includes the ability to identify emotions, to recognize their powerful effects, and to use that information to inform and guide behavior. Emotional intelligence includes the ability to influence–to evoke strong emotions in others, with a view to persuading or motivating them. It is more about focusing hard on both the person in front of you and your own emotions and reactions. 

What is artificial emotional intelligence?

Artificial emotional intelligence is the ability of the machine to recognize human emotions and then respond appropriately. The recognition and understanding of human emotions is crucial for AI systems to behave in appropriate ways according to the situation and smoothly integrate with all the different aspects of human life.  Currently, smartphones are used to allow voice assistants, like the iPhone’s Siri, to recognize and respond to user emotional concerns with appropriate information and supportive resources.

Strategies for Artificial Intelligence with Emotional Intelligence

Presently, AI is increasingly dependent on cloud computing, IoT and big data. Our goal is to model the range of higher human emotions, as well as their dynamics.  There are different frameworks, libraries, applications, toolkits, and datasets in the AI and machine learning world. By creating a direct neural interface with the Internet, humankind will be able to “plug into”  higher intelligence. The five components of AI with emotional intelligence are as follows; deep learning, self-awareness, safety and ethics, external awareness and big data collection and processing modules. Emotions  are essential part of human intelligence. Without emotional intelligence, AI is incomplete. Developing self-awareness of the machine is the first challenge of true AI based systems. 

Artificial Intelligence with Emotional Intelligence Issues and Challenges

Using artificial intelligence advances with emotional intelligence, there are several potential barriers to be addressed:

1. Privacy: Many people feel their emotions are private, and concerns about violations of privacy is genuine. Protective legislation will need to expand to include risks associated with AI, specifically the collection, storage, transfer and use of confidential health information.
2. Accuracy: AI accuracy in correctly determining emotional intent will need to be confirmed, specifically in regards to system biases or errors, before labeling a person as high or low emotional.
3. Safety: It is essential to ensure AI programs can appropriately respond to human users, so as to not worsen their emotional state or accidentally facilitate adverse situation.
4. Responsibility: Response protocols are needed on how to properly handle high risk cases that are flagged by AI technology, and what to do if AI risk assessments differ from human experts opinion.
5. Lack of understanding: There is a knowledge gap among key users on how AI technology fits into emotional understanding. More education on the topic is needed to address this.

Summary:

The criteria for what constitutes an artificial intelligent system has been shifted. Now developing emotional intelligence is one of the primary concern for AI research.  As technology is increasingly applied to situations where it must interact with human emotionally and intelligently. The most widely addressed area of research in automated emotion recognition, and where there has been the most progress, is the recognition of facial expressions. Physiological information has been shown to carry information that changes with different emotions. Handling of emotions, in others as well as in oneself, involves emotional intelligence. Integration of AI with emotional intelligence systems are expected to work alongside humans. The five components AI with emotional framework can provide some guidance in addressing this problem.

Sources:

1. Compassionate Artificial Superintelligence (AI-5.0) By Dr. Amit Ray, Inner Light Publishers, 2018.