Home » Talks
Category Archives: Talks
Artificial intelligence has learnt a lot from neuroscience. It was the move away from symbolic to neural net (machine learning) approaches that led to the current surge of interest in AI. Neural net approaches have enabled AI systems to do humanlike things such as object recognition and categorisation that had eluded the symbolic approaches.
So it was with great interest that I attended Dr. Tim Kietzmann's talk at the Cognitive and Brain sciences Unit (CBU) in Cambridge UK, earlier this month (March 2019), on what artificial intelligence (AI) and neuroscience can learn from each other.
Tim is a researcher and graduate supervisor at the MRC CBU and investigates principles of neural information processing using tools from machine learning and deep learning, applied to neuroimaging data recorded at high temporal (EEG/MEG) and spatial (fMRI) resolution.
Both AI and neuroscience aim to understand information processing and decision making - neuroscience primarily through empirical studies and AI primarily through computational modelling. The talk had symmetry. The first half was 'how can neuroscience benefit from artificial intelligence', and the second half was 'how artificial intelligence benefits from neuroscience'.
Types of AI
It is important to distinguish between 'narrow', 'general' and 'super' AI. Narrow AI is what we have now. In this context, it is the ability of a machine learning algorithm to recognise or classify particular things. This is often something visual like a cat or a face, but it could be a sound (as when an algorithm is used to identify a piece of music or in speech recognition).
General AI is akin to what people have. When or if this will happen is speculative. Ray Kurzweil, Google’s Director of Engineering, predicts 2029 as the date when an AI will pass the Turing test (i.e. a human will not be able to tell the difference between a person and an AI when performing tasks). The singularity (the point when we will multiply our effective intelligence a billion fold by merging with the intelligence we have created), he predicts should happen by about 2045. Super AIs exceed human intelligence. Right now, they only appear in fiction and films.
It is impossible to predict how this will unfurl. After all, you could argue that the desktop calculator several decades ago exceeded human capability in the very narrow domain of performing mathematical calculations. It is possible to imagine many very narrow and deep skills like this becoming fully integrated within an overall control architecture capable of passing results between them. That might look quite different from human intelligence.
One Way or Another
Research in machine learning, a sub-discipline of AI, has given neuroscience researchers pattern recognition techniques that can be used to understand high-dimensional neural data. Moreover, the deep learning algorithms, that have been so successful in creating a new range of applications and interest in AI, offer an exciting new framework for researchers like Tim and colleagues, to advance knowledge of the computational principles at play in the brain. AI allows researchers to test different theories of brain computations and cognitive function by implementing and testing them. 'Today's computational neuroscience needs machine learning techniques from artificial intelligence'.
AI benefits from neuroscience by informing the development of a wide variety of AI applications from care robots to medical diagnosis and self-driving cars. Some principles that commonly apply in human learning (such as building on previous knowledge and unsupervised learning) are not yet integrated into AI systems.
For example, a child can quickly learn to recognise certain types of objects, even those such as a mythical 'Tufa' that they have never seen before. A machine learning algorithm, by contrast, would require tens of thousands of training instances in order to reliably perform that same task. Also, AI systems can easily be fooled in ways that a person never would. Adding a specially crafted 'noise' to an image of a dog, can lead an AI to misclassify it as an ostrich. A person would still see a dog and not make this sort of mistake. Having said that, children will over-generalise from exposure to a small number of instances, and so also make mistakes.
It could be that the column structures found in the cortex have some parallels to the multi-layered networks used in machine learning and might inform how they are designed. It is also worth noting that the idea of reinforcement learning used to train artificial neural nets, originally came out of behavioural psychology - in particular Pavlov and Skinner. This illustrates the 'intertwined' nature of all these disciplines.
The Neuroscience of Ethics
Although this was not covered in the talk, when it comes to ethics, neuroscience may have much to offer AI, especially as we move from narrow AI into artificial general intelligence (AGI) and beyond. Evidence is growing as to how brain structures, such as the pre-frontal cortex are involved in inhibiting thought and action. Certain drugs affect neuronal transmission and can disrupt these inhibitory signals. Brain lesions and the effects of strokes can also interfere with moral judgements. The relationship of neurological mechanisms to notions of criminal responsibility may also reveal findings relevant to AI. It seems likely that one day the understanding of the relationship between neuroscience, moral reasoning and the high-level control of behaviours will have an impact on the design of, and architectures for, artificial autonomous intelligent systems (i.e. see Neuroethics: Challenges for the 21st Century.Neil Levy - 2007 - Cambridge University Press or A Neuro-Philosophy of Human Nature: Emotional Amoral Egoism and the Five Motivators of Humankind - April 2019).
Understanding the Brain
The reality of the comparison between human and artificial intelligence comes home when you consider the energy requirements of the human brain and computer processors performing similar tasks. While the brain uses about 15 watts of energy, just a single graphics processing unit requires up to 250 watts.
It has often been said that you cannot understand something until you can build it. That provides a benchmark against which we can measure our understanding of neuroscience. Building machines that perform as well as humans is a necessary step in that understanding, although that still does not imply that the mechanisms are the same.
Algorithms can determine whether you get a loan, predict what diseases you might get and even assess how long you might live. It’s kind of important we can trust them!
David Spiegelhalter is the Winton Professor for the Public Understanding of Risk in the Statistical Laboratory, Centre for Mathematical Sciences at the University of Cambridge. As part of the Cambridge Science Festival he was talking (21st of March 2019) on the subject of making algorithms trustworthy.
I’ve heard David speak on many occasions and he is always informative and entertaining. This was no exception.
Algorithms now regularly advise on book and film recommendations. They work out the routes on your satnav. They control how much you pay for a plane ticket and, annoyingly, they show you advertisements that seem to know far too much about you.
But more importantly they can affect life and death situations. The results of an algorithmic assessment of what disease you might have could be highly influential, affecting your treatment, your well-being and on your future behaviour.
David is a fan of Onora O’Neill who suggests that organisations should not be aiming to increase trust but should aim to demonstrate trustworthiness. False claims about the accuracy of algorithms are as bad as defects in the algorithms themselves.
The pharmaceutical industry has long used a phased approach assessing the effectiveness, safety and side-effects of drugs. This includes the use of randomly controlled trials, and long-term surveillance after a drug comes onto the market, to spot rare side-effects.
The same sorts of procedures should be applied to algorithms. However, currently only the first phase of testing on new data is common. Sometimes algorithms are tested against the decisions that human experts make. Rarely will randomly controlled trials be conducted, or the algorithm in use be subject to long-term monitoring.
Algorithms should be transparent. They should be able to explain their decisions as well as to provide them. But transparency is not enough. O’Neill uses the term 'intelligent openness’ to describe what is required. Explanations need to be accessible, intelligible, usable, and assessable.
Algorithms need to be both globally and locally explainable. Global explainability relates to the validity of the algorithm in general, while local explainability relates to how the algorithm arrived at a particular decision. One important way of being able to test an algorithm, even when it’s a black box, is to be able to play with inputting different parameters and seeing the result.
Deep Mind (owned by Google) is looking at how explanations can be generated from intermediate stages of the operation of machine learning algorithms.
Explanation can be provided at many levels. At the top level this might be a simple verbal summary. At the next level it might be having access to a range of graphical and numerical representations with the ability to run 'what if' queries. At a deeper level, text and tables might show the procedures that the algorithm used. Deeper still, would be the mathematics underlying the algorithm. Lastly, the code that runs the algorithm should be inspectable. I would say that a good explanation is dependent on understanding what the user wants to know - in other words, it is not just a function of the decision making process but also a function of the user’s actual and desired state of knowledge.
Without these types of explanation, algorithms such as the one used by the US company Compas to predict rates of recidivism, are difficult to trust.
It is easy to feel that an algorithm is unfair or can’t be trusted. If it cannot provide sufficiently good explanations, and claims about it are not scientifically substantiated, then it is right to be sceptical about its decisions.
Most of David’s points apply more broadly than to artificial intelligence and robots. They are general principles applying to the transparency, accountabilityand user acceptance of any system. Trust and trustworthiness are everything.
See more of David’s work on his personal webpage at http://www.statslab.cam.ac.uk/Dept/People/Spiegelhalter/davids.html , . And read his new book “The Art of Statistics: Learning from Data”, available shortly.
Cambridge (UK) is awash with talks at the moment, and many of these are about artificial intelligence. On Tuesday (12th of March 2019) I went to a talk, as part of Cambridge University’s science festival, by José Hernández-Orallo (Universitat Politècnica de València), titled Natural or 'Artificial Intelligence? Measures, Maps and Taxonomies'.
José opened by pointing out that artificial intelligence was not a subset of human intelligence. Rather, it overlaps with it. After all, some artificial intelligence already far exceeds human intelligence in narrow domains such as playing games (Go, Chess etc.) and some identification tasks (e.g. face recognition). But, of course, human intelligence far outstrips artificial intelligence in its breadth and the amount of training needed to learn concepts.
José‘s main message was how, when it comes to understanding artificial intelligence, we (like the political scene in Britain at the moment) are in uncharted territory. We have no measures by which we can compare artificial and human intelligence or to determine the pace of progress in artificial intelligence. We have no maps that enable us to navigate around the space of artificial intelligence offerings (for example, which offerings might be ethical and which might be potentially harmful). And lastly, we have no taxonomies to classify approaches or examples of artificial intelligence.
Whilst there are many competitions and benchmarks for particular artificial intelligence tasks (such as answering quiz questions or more generally reinforcement learning), there is no overall, widely used classification scheme.
My own take on this is to suggest a number of approaches that might be considered. Coming from a psychology and psychometric testing background, I am aware of the huge number of psychological testing instruments for both intelligence and many other psychological traits. See for example, Wikipedia or the British Psychological Society list of test publishers. What is interesting is that, I would guess, most software applications that claim to use artificial intelligence would fail miserably on human intelligence tests, especially tests of emotional and social intelligence. At the same time they might score at superhuman levels with respect to some very narrow capabilities. This illustrates just how far away we are from the idea of the singularity - the point at which artificial intelligence might overtake human intelligence.
Another take on this would be to look at skills. Interestingly, systems like the Amazon's Alexa describe the applications or modules that developers offer as 'skills'. So for example, a skill might be to book a hotel or to select a particular genre of music. This approach defines intelligence as the ability to effectively perform some task. However, by any standard, the skill offered by a typical Alexa 'skill', Google Home or Siri interaction is laughably unintelligent. The artificial intelligence is all in the speech recognition, and to some extent the speech production side. Very little of it is concerned with the domain knowledge. Even so, a skills based approach to measurement, mapping and taxonomy might be a useful way forward.
When it comes to Ethics, There are also some pointers to useful measures, maps and taxonomies. For example the blog post describing Josephine Young’s work identifies a number of themes in AI and data ethics. Also, the video featuring Dr Michael Wilby on the http://www.robotethics.co.uk/robot-ethics-video-links/ page starts with a taxonomy of ethics and then maps artificial intelligence into this framework.
But, overall, I would agree with José that there is not a great deal of work in this important area and that it is ripe for further research. If you are aware of any relevant research then please get in touch.
John Wyatt is a doctor, author and research scientist. His concern is the ethical challenges that arise with technologies like artificial intelligence and robotics. On Tuesday this week (11th March 2019) he gave a talk called ‘What does it mean to be human?’ at the Wesley Methodist Church in Cambridge.
To a packed audience, he pointed out how interactions with artificial intelligence and robots will never be the same as the type of ‘I – you’ relationships that occur between people. He emphasised the important distinction between ‘beings that are born’ and ‘beings that are made’ and how this distinction will become increasingly blurred as our interactions with artificial intelligence become commonplace. We must be ever vigilant against the use of technology to dehumanise and manipulate.
I can see where this is going. The tendency for people to anthropomorphise is remarkably strong - ‘the computer won’t let me do that’, ‘the car has decided not to start this morning’. Research shows that we can even attribute intentions to animated geometrical shapes ‘chasing’ each other around a computer screen, let alone cartoons. Just how difficult is it going to be to not attribute the ‘human condition’ to a chatbot with an indistinguishably human voice or a realistically human robot. Children are already being taught to say ‘please’ and ‘thank you’ to devices like Alexa, Siri and Google Home – maybe a good thing in some ways, but …
One message I took away from this talk was a suggestion for a number of new human rights in this technological age. These are: (1) The right to cognitive liberty (to think whatever you want), (2) The right to mental privacy (without others knowing) (3) The right to mental integrity and (4) The right to psychological continuity - the last two concerning the preservation of ‘self’ and ‘identity’.
A second message was to consider which country was most likely to make advances in the ethics of artificial intelligence and robotics. His conclusion – the UK. That reassures me that I’m in the right place.
See more of John’s work, such as his essay ‘God, neuroscience and human identity’ at his website johnwyatt.com