Written essay

Your project group will write an essay about the historical, ethical and/or philosophical aspects of an AI topic. The essay can be expository, analytical or argumentative, or all of them. A list of suggested topics with suggested reading material is available below. You can also come up with a topic of your own – ask your supervisor in that case.

You submit your essay via the conference management system EasyChair. Then you will read and review essays by two other groups. Finally, you get a chance to incorporate feedback from the reviews and re-submit your essay.

Organisation

Claes Strannegård is in charge of the essay organisation and will be supervising your essays. He has a background in mathematical logic, AI-entrepreneurship, and more recently in artificial general intelligence research.

As you know you will be working with your essays in the groups that you have already formed. Here is a description the workflow of the writing process and give you the relevant deadlines.

As a first step you should try to come up with a suitable topic. The topic should be interesting to everyone in the group and concern philosophical, historical, or ethical aspects of AI.

Many examples of essay topics can be found on the web page of the course. You are very welcome to come up with your own topic too, but in that case you have to contact Claes via email, so that he can confirm that it is ok.

Each group will have an initial 30 minute meeting with Claes on one of the following days:

Please book a 30-minute slot here (one per group, write your group name in the “name” field):

At this first meeting we will discuss the outline and the content of your essays. Please try to be well underway with your work when we meet, so that the meeting can be as productive as possible.

The deadline for uploading the first draft of the essay is May 12. Upload it to Easychair as described on the course webpages.

Your next task is to review two essays. The deadline for uploading your two reviews to Easychair is May 19.

So, from May 20 you will be able to see the feedback on your first draft from two student groups and also from Claes on Easychair.

Next you go to work with the final version of your essay. There will be drop-in sessions as follows when you can discuss your essays with Claes:

We expect you to come prepared with specific questions to those drop-in sessions.

The deadline for uploading the final version to Easychair is June 2.

You can contact Claes via email or phone if you run into trouble with your essay, or if you have clarification questions, etc. We will then try to solve the issue in the most suitable manner. If there are issues with the schedule for your group, please contact Claes in the first week of April.

Good luck with your essays!

Guidelines

Formatting

Your report should adhere to the official style guidelines for the Coling 2016 conference. Download the following zip file which contains LaTeX files, a Microsoft Word template file and a sample PDF file:

The PDF file contains a description of the style guidelines. Everything that is written in the guidelines is to be followed, except for the following details:

In particular, remember the following:

If you want examples of what a research paper should look like, there are lots of different open repositories, such as the ACL Anthology. (See, e.g., the proceedings of COLING 2016).

Writing guidelines

An expository essay is an explanation of a specific idea. An analytical essay is a review (your interpretation) of existing literature (books/articles) on a topic. Finally, an argumentative essay is more like an opinion piece.

An argumentative essay is based on arguing your opinion on a debated subject, but the other two essay types also allow you to have an opinion on the matter at hand. Specifically, an analytical essay benefits from having a clear thesis about e.g. what the author you’re reviewing meant, what views they’re expressing, a movement they exemplify or something similar.

When drawing from books and articles, consider the possibility of presenting two different viewpoints, presenting both sides of an argument.

The essay, taken on its own, should be clear and self-contained. You should not require that the reviewers read the articles you are addressing in order to understand it. This should guide your writing process most of the time. Nevertheless, be prepared that the some reviewers might actually read your references! This means, of course, that you shouldn’t misrepresent the sources you’re citing—but also that you should not bore readers who have read a referenced article by repeating the whole thing.

Irrespective of the type of essay you choose to write, remember the following things:

Reviewing guidelines

For each of the essays you review, there will be a standard form (provided through EasyChair) in which you paste the raw text of your review. This means the reviews need not be formatted, apart from paragraph breaks and possibly headers. Each essay is scored based on certain criteria, and you can provide subjective comments to elaborate your scores or ask for clarifications.

Before you start grading an essay written by a different group, try to establish:

Does the essay provide enough support for their perspective while aligning with or against the original author(s)? Is the reasoning sound and acceptable to you? Try to apply the grading criteria described above (clarity, soundness, originality, previous work).

Finally, the following can be treated as good practices during peer-reviewing:

Don’t make the reviews too long! I think EasyChair has a limit of 2000 chars, or 500 words or something – look, it doesn’t matter, the point is you’re not writing a new paper here.

Submission

You submit your essay in PDF format using the EasyChair conference system:

First, one of you has to create an EasyChair account, and then that person becomes the “corresponding author”. (Who this is has no effect whatsoever on your grades).

After the submission deadline, you will be assigned 2 essays written by your fellow students which you should read and review. You will submit your review via EasyChair too. After the review process, each group will be able to see the reviews that they got for their own essays.

After you create an account on EasyChair, please notify Claes (and your supervisor) all details including the email address used to signup on EasyChair, your group number and name. Also include the topic you have chosen to write your essay on. Do not forget this as you will not be able to review essays without doing this.

Final essay

You have to revise your essay according to the reviews you received. Note that you do not have to follow every advice that the reviewers gave, only those that make sense to you and that you believe will make your essay better.

Important: Together with your final essay, you have to write a short text describing which of the reviewers’ issues you have addressed (and in which way) and what you have decided not to care about. Add this description as an extra page at the end of your essay (as an appendix, after the references).

Grading

The grades are set by us teachers, not by your fellow students. Nevertheless, the reviews you get will influence your final grade. But so will the reviews you write! We will read all essays and all reviews, and compare them with each other when setting the final grades.

Your essay can receive 0–3 bonus points counting towards a higher grade. Our assessment of the essays is based on the following four criteria:

Clarity: A good essay has a correct and clear use of language and text structure, facilitating the reader’s understanding. Avoid vagueness and hand-waving, and jumping between points.

Soundness: In a good essay, the views and conclusions presented in the essay are well supported. (Remember, skepticism is also a viewpoint, and “let’s wait and see” is also a conclusion.) Avoid unfounded assumptions and assertions, and question them if they appear in the materials you draw from.

Originality: A good essay includes the writers’ own intellectual contributions. (In this setting you will not do original research, but arguing your own viewpoint, and synthesising available information into a clear explanation, both count.) Avoid repeating what others have said if you don’t have anything to add or comment.

Building on previous work: A good essay fits in the context of the current state of the field and public discourse on the subject. Avoid writing an essay incorporating only information from Wikipedia and your own private musings. (Such an essay might be interesting, but we ask for more here.)

When it comes to the reviews, the same criteria apply, to some extent. The most important one is clarity: a good review constructively presents concrete suggestions for improvement, in a clear and precise fashion. Your group can get 0–1 bonus points for your two reviews; the bonus point is awarded only if both your reviews are good.

Topics

Below are some essay topics that you might want to consider. Each topic includes a selection of relevant reading material that might be useful as starting points, sources of inspiration and/or information for the essay. The material for each topic is by no means exhaustive, and you are welcome to look for other sources if you want to. You might also find some overlap in the suggested material for different topics, since some of them are closely related.

The suggested topics are not set in stone: you could use them as starting points, and combine or alter them if you wish. For example, you could turn one of the expository essays into an analytical one on the same subject.

When you have decided on a topic, whether it is one of the suggestions or not, notify your supervisor. If it’s a new topic, they’ll have to okay it before you start writing, but most likely they will encourage your unique choice!

A note: the suggested topics vary between general and more specific. Thus, remember that the essay doesn’t have to be on a “big idea” in AI ethics, philosophy or history, even though it can be. Other options include a specific technological advance with impact on AI, or the impact of AI an another field such as medicine, language technology etc.

The history of artificial neural networks in AI

Give an account of AANs, from the early years in the 1940s to current state of the art. Explain how people have thought about them in different times, their successes and failures. You don’t have to go into detail about the mathematics behind ANNs, but you should be able to give a simple classification of different types of networks, and when and why they boomed and crashed.

Material:

  1. Chapter 1 from A Brief Introduction to Neural Networks by David Kriesel
  2. Artificial neural network (Wikipedia)
  3. History of artificial intelligence
  4. Neural Networks by Christos Stergiou and Dimitrios Siganos

The state of the art in video-game AI

Write a survey of current video-game AI techniques, classifying them and comparing them with previous approaches and noting the breakthroughs that have been made in this field. In particular, you could discuss the challenges that arise in this context, such as the limitations of reasoning under uncertainty and the real-time constraints usually placed on video game AI.

Material:

  1. Artificial intelligence (video games) (Wikipedia)
  2. Call for AI Research in RTS Games, by Michael Buro
  3. State-Driven Game Agent Design
  4. When artificial intelligence in video games becomes…artificially intelligent by Kris Graft
  5. Artificial Intelligence in Games by James Wexler
  6. Game AI: The State of the Industry by Steven Woodcock

The ethics of AI (machine ethics)

Write an essay presenting and discussing the ethical issues surrounding AI. This can include specific AI (self-driving cars) or general AI (human-level cognition and consciousness). Here are some questions that you might want to consider: Can we give AI a sense of right and wrong? If an AI harms humans, who is to blame? Would a sentient AI have the same rights as humans?

Material:

  1. Roboethics: Social and Ethical Implications of Robotics by Gianmarco Veruggio and Fiorella Operto
  2. The Ethics of Artificial Intelligence by Nick Bostrom
  3. Ethical Guidelines for A Superintelligence by Ernest Davis
  4. Why you shouldn’t worry about liability for self-driving car accidents by Mark Harris

The history of expert systems in AI

Give an account of the history of expert systems, from their inception in the 1970s to their apex in the late 80s and their subsequent downfall. Discuss how the ideas behind these systems arose and how suitable they have been in the context of decision making. You should be able to pinpoint success stories and failures, citing the relevant literature. You may also want to tackle the question of whether expert systems as a whole have utterly failed or if their core ideas have just been assimilated into other fields (considering this a form of success).

Material:

  1. Expert system (Wikipedia)
  2. IBM’s Watson: Has the Time Come for Expert Systems in Medicine?
  3. The Structure of MYCIN System (Book chapter) by William van Melle
  4. Expert Systems

What will happen when machines can think?

Write a philosophical essay exploring a hypothetical future when machines are able to think and perform cognitive tasks with the same proficiency as human beings, with a focus on predicting these events. Can intelligent machines be considered conscious? This is expected to be mostly an opinion piece, though you can quote other authors to compare and contrast their ideas as long as you give them proper credit.

Material:

  1. Enthusiasts and Skeptics Debate Artificial Intelligence by Kurt Andersen
  2. Hyping Artificial Intelligence – Yet Again by Gary Marcus
  3. How We’re Predicting AI—or Failing To by Stuart Armstrong and Kaj Sotala
  4. How Long Before Superintelligence? by Nick Bostrom
  5. Future Progress in Artificial Intelligence: A Survey of Expert Opinion by Vincent C. Muller and Nick Bostrom

The economic impact of AI: will humans ever become redundant?

Study how the development of AI will affect economy and human society. What will humans do if AI could take care of all current human jobs? Maybe then humans could just do leisure and social activities all the time instead of having to work. Discuss the pros and cons of this scenario, as well as the likelihood of this ever happening. This can be either a review, a literature survey or an opinion piece.

Material:

  1. Humans need not apply, by C.G.P.Grey
  2. Why the future doesn’t need us by Bill Joy
  3. AI, Robotics and the Future of Jobs by Aaron Smith and Janna Anderson
  4. Artificial Intelligence, Employment and Income by Nils J. Nilsson
  5. The Impacts of Robotics, Artificial Intelligence On Business and Economics by Cüneyt Dirican

How close are we to Technological Singularity?

Write an essay presenting the concept of singularity, how it has evolved over the years and where it is currently going. Discuss the potential impact, both good and bad, of a scenario where AI can get exponentially better at recursive self-improvement, ultimately leading to superintelligence. Summarise the views of scientists and futurists about whether this is a real possibility in the near future or at all (and give your own opinion if you like).

Material:

  1. Ten Years to the Singularity: If We Really, Really Try by Ben Goertzel.
  2. The Singularity Myth by Theodore Modis.
  3. The Singularity Isn’t Near by Paul Allen.
  4. The Coming Technological Singularity: How to Survive in the Post-Human Era by Vernor Vinge

Revisiting the Turing Test – is it good enough?

Analyse the suitability of the Turing Test. You could include a survey of past attempts at setting up or passing the Turing Test, such as chatbot competitions. Is it a good indicator of intelligence? Is there any better alternative? You can summarise the opinions of experts on this topic and/or provide your own.

Material:

  1. Computing Machinery and Intelligence by Alan Turing
  2. Lessons from a Restricted Turing Test by Stuart M. Shieber
  3. Minds, Brains, and Programs by Searle, John R
  4. Turing tested: An interview with Eugene Goostman, the first computer programme to pass for human by Chris Green

A survey of machine learning techniques

Give an account of machine learning techniques, including the several approaches that have been tried in the history of the field and the cutting-edge applications. Compare machine learning with other approaches to AI, such as expert systems. Classify machine learning techniques and discuss real-world applications, e.g. natural language processing, OCR, image recognition, etc.

Material:

  1. Machine learning (Wikipedia)
  2. The wonderful and terrifying implications of computers that can learn, a TED talk by Jeremy Howard
  3. IBM Watson
  4. A Few Useful Things to Know about Machine Learning
  5. Chapter 1 of Introduction to Machine Learning by Ethem Alpaydın

The history of board-game AI

Write an essay about AI for playing board games. Give an account of the history of the field, marking its successes and failures. You can either focus on a particular game or give a general overview. You should be able to explain and classify different techniques (without going into too many details) and/or what techniques are more appropriate for different games.

Material:

  1. Computer chess (Wikipedia)
  2. Game Over – Kasparov And The Machine
  3. Computer Go (wikipedia)
  4. TD-Gammon
  5. Monte-Carlo Tree Search: A New Framework for Game AI by Guillaume Chaslot, Sander Bakkes, Istvan Szita and Pieter Spronck
  6. Computers Solve Checkers—It’s a Draw by JR Minkel

Autonomous vehicles

Give an account of the state of the art in autonomous vehicles. Provide a general description of how self-driving cars work, show the practical tests that have been carried out and discuss the societal issues that come about from the introduction of such vehicles. For example, who should be to blame for an accident? A huge part of all human jobs consist in the transport of goods; what will those people do if trucks can drive themselves?

Material:

  1. How Google’s self-driving car works by Erico Guizzo
  2. Chapter 2 of Autonomous Vehicle Technology by RAND Corporation
  3. Autonomous car (Wikipedia)
  4. Why you shouldn’t worry about liability for self-driving car accidents by Mark Harris
  5. Statistics about jobs in the transport sector (US)
  6. Where to? A History of Autonomous Vehicles by Marc Weber

Can machines turn against us? Would AI be beneficial to humanity as a whole?

Write a philosophical essay concerning the dichotomy between friendly and evil AI. Discuss the problems that might arise from a non-friendly AI which might see no need to keep humans around. Should we stop developing AI as a whole? Consider ways of limiting the impact of AI or countering hostile AIs.

Material:

  1. Artificial Intelligence as a Positive and Negative Factor in Global Risk by Eliezer Yudkowsky
  2. Omohundro’s “Basic AI Drives” and Catastrophic Risks by Carl Shulman
  3. Artificial Intelligence: Past and Future by Moshe Y. Vardi
  4. Our Final Invention: Artificial Intelligence and the End of the Human Era (Book) by James Barrat
  5. Here Be Dragons: Science, Technology and the Future of Humanity (Book) by Olle Häggström

Artificial Creativity

Give an account of Artificial Creativity, the field which attempts to teach machines to be creative, usually in an artistic sense. These creative activities might involve composing music, painting, writing poetry, etc; you are free to chose to focus on a specific form of creativity or to discuss all of them. Provide a general overview of the field, and include references to previous and current approaches, comparing their relative merits.

Material:

  1. Computational creativity (Wikipedia)
  2. Computers at the dawn of creativity
  3. Can Computers Be Creative?
  4. Emily Howell (composer AI)

The history of machine translation

Write a survey of the historical development of the field of machine translation, i.e. software that translates from one natural language to another. Speech-recognition and synthesis, though not essential to the translation itself, are found in combination in some systems to make them more usable. Classify the different approaches to machine translation, from traditional rule-based methods to statistical methods. Compare and contrast them and study how successful they have been in practice.

Material:

  1. Translation by Warren Waever, written in 1949!
  2. Machine translation (Wikipedia)
  3. Approaches to machine translation by Sneha Tripathi and Juran Krishna Sarkhel
  4. Machine translation: a concise history by W. John Hutchins

Additional material

Here you can find additional material for your essay. Everything that follows should be considered optional material.

First a couple of inspirational videos:

And here is an unsorted list of articles and books that might be of interest to you:

Title Author Type
The Society of Mind Marvin Minsky Book
Computer Power and Human Reason: From Judgment To Calculation Joseph Weizenbaum Book
Teknik och etik (in Swedish) Sven Ove Hansson Book
IBM’s Watson takes on brain cancer Ajay Royyuru Article
Sir Arthur C. Clarke’s Predictions Sir Arthur C. Clarke Article
Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. A. Garg et. al Article
Rule-Based Expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming Project Bruce G. Buchanan, Edward H. Shortliffe Book
Computer Power and Human Reason Joseph Weizenbaum Book
Alchemy and AI Hubert Dreyfus Book
What Computers Can’t Do Hubert Dreyfus Book
Moral Machines: Teaching Robots Right from Wrong Wendell Wallach, Colin Allen Book
Mind over Machine Hubert Dreyfus Book
The Singularity Is Near: When Humans Transcend Biology Ray Kurzweil Book
Superintelligence: Paths, Dangers, Strategies Nick Bostrom Book
The Quest for Artificial Intelligence Nils J. Nilsson Book
How the Mind Works Steven Pinker Book
Being-in-the-World: A Commentary on Heidegger’s Being and Time Hubert Dreyfus Book
One Half a Manifesto Jaron Lanier Interview
Transcending Complacency on Superintelligent Machines Stephen Hawking Article
Smarter Than Us: The Rise of Machine Intelligence Stuart Armstrong Book
Machines Who Think: A Personal Inquiry Into the History and Prospects of Artificial Intelligence Pamela McCorduck Book
Gödel, Escher, Bach: An Eternal Golden Braid Douglas Hofstadter Book