Alignment Problem

The Alignment Problem: Machine Learning and Human Values
by Brian Christian
Published 5 Oct 2020

Norton Special Sales at specialsales@wwnorton.com or 800-233-4830 Jacket design: Gregg Kulick Jacket illustrations: (top) Courtesy of Legado Cajal, Instituto Cajal (CSIC), Madrid; (bottom) Gregg Kulick Book design by Chris Welch Production manager: Lauren Abbate The Library of Congress has cataloged the printed edition as follows: Names: Christian, Brian, 1984– author. Title: The alignment problem : machine learning and human values / Brian Christian. Description: First edition. | New York, NY : W.W. Norton & Company, [2020] | Includes bibliographical references and index. Identifiers: LCCN 2020029036 | ISBN 9780393635829 (hardcover) | ISBN 9780393635836 (epub) Subjects: LCSH: Artificial intelligence—Moral and ethical aspects. | Artificial intelligence—Social aspects. | Machine learning—Safety measures. | Software failures. | Social values. Classification: LCC Q334.7 .C47 2020 | DDC 174/.90063—dc23 LC record available at https://lccn.loc.gov/2020029036 W.

…

Ghahramani, Zoubin. “Probabilistic Machine Learning and Artificial Intelligence.” Nature 521, no. 7553 (2015): 452–59. Ghorbani, Amirata, Abubakar Abid, and James Zou. “Interpretation of Neural Networks Is Fragile.” In Proceedings of the AAAI Conference on Artificial Intelligence 33 (2019): 3681–88. Gielniak, Michael J., and Andrea L. Thomaz. “Generating Anticipation in Robot Motion.” In 2011 RO-MAN, 449–54. IEEE, 2011. Giusti, Alessandro, Jérôme Guzzi, Dan C. Cireşan, Fang-Lin He, Juan P. Rodríguez, Flavio Fontana, Matthias Faessler, et al. “A Machine Learning Approach to Visual Perception of Forest Trails for Mobile Robots.”

…

Liu, Lydia T., Sarah Dean, Esther Rolf, Max Simchowitz, and Moritz Hardt. “Delayed Impact of Fair Machine Learning.” In Proceedings of the 35th International Conference on Machine Learning, 2018. Liu, Si, Risheek Garrepalli, Thomas G. Dietterich, Alan Fern, and Dan Hendrycks. “Open Category Detection with PAC Guarantees.” In Proceedings of the 35th International Conference on Machine Learning, 2018. Liu, Si, Risheek Garrepalli, Alan Fern, and Thomas G. Dietterich. “Can We Achieve Open Category Detection with Guarantees?” In Workshops at the Thirty-Second AAAI Conference on Artificial Intelligence, 2018. Lockhart, Ted. Moral Uncertainty and Its Consequences.

Architects of Intelligence
by Martin Ford
Published 16 Nov 2018

About a quarter of those interviewed in this book are women, and that number is likely significantly higher than what would be found across the entire field of AI or machine learning. A recent study found that women represent about 12 percent of leading researchers in machine learning. (https://www.wired.com/story/artificial-intelligence-researchers-gender-imbalance) A number of the people I spoke to emphasized the need for greater representation for both women and members of minority groups. As you will learn from her interview in this book, one of the foremost women working in artificial intelligence is especially passionate about the need to increase diversity in the field. Stanford University’s Fei-Fei Li co-founded an organization now called AI4ALL (http://ai-4-all.org/) to provide AI-focused summer camps geared especially to underrepresented high school students.

…

In recent years, he has focused on AI and machine learning and worked on the development of TensorFlow, Google’s widely-used open source software for deep learning. He currently guides Google’s future path in AI as director of artificial intelligence and head of the Google Brain project. MARTIN FORD: As the Director of AI at Google and head of Google Brain, what’s your vision for AI research at Google? JEFF DEAN: Overall, I view our role as to advance the state of the art in machine learning, to try and build more intelligent systems by developing new machine learning algorithms and techniques, and to build software and hardware infrastructure that allows us to make faster progress on these approaches and allow other people to also apply these approaches to problems they care about.

…

One especially important audience, however, consists of young people who might consider a future career in artificial intelligence. There is currently a massive shortage of talent in the field, especially among those with skills in deep learning, and a career in AI or machine learning promises to be exciting, lucrative and consequential. As the industry works to attract more talent into the field, there is widespread recognition that much more must be done to ensure that those new people are more diverse. If artificial intelligence is indeed poised to reshape our world, then it is crucial that the individuals who best understand the technology—and are therefore best positioned to influence its direction—be representative of society as a whole.

Possible Minds: Twenty-Five Ways of Looking at AI
by John Brockman
Published 19 Feb 2019

Instead they take in information from others in a remarkably subtle and sensitive way, making complex inferences about where the information comes from and how trustworthy it is and systematically integrating their own experiences with what they are hearing.* “Artificial intelligence” and “machine learning” sound scary. And in some ways they are. These systems are being used to control weapons, for example, and we really should be scared about that. Still, natural stupidity can wreak far more havoc than artificial intelligence; we humans will need to be much smarter than we have been in the past to properly regulate the new technologies. But there is not much basis for either the apocalyptic or the utopian vision of AIs replacing humans. Until we solve the basic paradox of learning, the best artificial intelligences will be unable to compete with the average human four-year-old.

…

If they can’t infer what we value, there’s no way for them to act in support of those values—and they may well act in ways that contravene them. Value alignment is the subject of a small but growing literature in artificial-intelligence research. One of the tools used for solving this problem is inverse-reinforcement learning. Reinforcement learning is a standard method for training intelligent machines. By associating particular outcomes with rewards, a machine-learning system can be trained to follow strategies that produce those outcomes. Wiener hinted at this idea in the 1950s, but the intervening decades have developed it into a fine art. Modern machine-learning systems can find extremely effective strategies for playing computer games—from simple arcade games to complex real-time strategy games—by applying reinforcement-learning algorithms.

…

But, at least for now, we have almost no idea at all how the sort of creativity we see in children is possible.” Everyone’s heard about the new advances in artificial intelligence, and especially machine learning. You’ve also heard utopian or apocalyptic predictions about what those advances mean. They have been taken to presage either immortality or the end of the world, and a lot has been written about both of those possibilities. But the most sophisticated AIs are still far from being able to solve problems that human four-year-olds accomplish with ease. In spite of the impressive name, artificial intelligence largely consists of techniques to detect statistical patterns in large data sets.

The Myth of Artificial Intelligence: Why Computers Can't Think the Way We Do
by Erik J. Larson
Published 5 Apr 2021

Classification: LCC Q 335 .L37 2021 | DDC 006.3—dc23 LC record available at https://lccn.loc.gov/2020050249 To Brooke and Ben CONTENTS Introduction Part I: THE SIMPLIFIED WORLD   1   The Intelligence Error   2   Turing at Bletchley   3   The Superintelligence Error   4   The Singularity, Then and Now   5   Natural Language Understanding   6   AI as Technological Kitsch   7   Simplifications and Mysteries Part II: THE PROBLEM OF INFERENCE   8   Don’t Calculate, Analyze   9   The Puzzle of Peirce (and Peirce’s Puzzle) 10   Problems with Deduction and Induction 11   Machine Learning and Big Data 12   Abductive Inference 13   Inference and Language I 14   Inference and Language II Part III: THE FUTURE OF THE MYTH 15   Myths and Heroes 16   AI Mythology Invades Neuroscience 17   Neocortical Theories of Human Intelligence 18   The End of Science? NOTES ACKNOWLEDGMENTS INDEX INTRODUCTION In the pages of this book you will read about the myth of artificial intelligence. The myth is not that true AI is possible. As to that, the future of AI is a scientific unknown. The myth of artificial intelligence is that its arrival is inevitable, and only a matter of time—that we have already embarked on the path that will lead to human-level AI, and then superintelligence.

…

It’s impossible to make sense of language relying on induction, correctly understood (that is, not packing into it other forms of inference). 14. Marcus and Davis, Rebooting AI. 15. Russell, Human Compatible. 16. Pearl, The Book of Why, 36. Chapter 11: Machine Learning and Big Data 1. Stuart Russell, Human Compatible: Artificial Intelligence and the Problem of Control (New York: Viking, 2019). 2. Tom Mitchell, Machine Learning (New York: McGraw-Hill Education, 1997), 2. 3. We trust the filters in large part because they are deliberately permissive: spam is more likely to get into inboxes then legitimate messages are to get thrown out.

…

computer (Watson) by, 220–226, 230–231 ImageNet competitions, 135, 145, 155, 165, 243 image recognition, 278–279 imitation game, 9, 51 The Imitation Game (film), 21 incompleteness theorems, 12–15 induction, 115–121, 171–172; abduction and, 161; in artificial intelligence, 273–274; in life situations, 125–126; limits to, 278–279; machine learning as, 133; not strategy for artificial general intelligence, 173; problems of, 122–124; regularity in, 126–129 inductive inference, 189 inference, 4, 104, 280–281, 283n1; abductive inference, 99–102, 162–163; in artificial intelligence, 103; combining types of, 218–219, 231; guesses as, 160; in history of science, 103–105; in knowledge bases, 181, 182; monotonic, 167; non-monotonic, 167–168; as trust, 129–130; types of, 171 inference engines, 182–184 information technology, 249, 252 ingenuity: Gödel on, 12; Turing on, 11, 17, 18 innovation: decline in, 269; funding in control over, 270 insight, 103 instincts, 184 intelligence.

Literary Theory for Robots: How Computers Learned to Write
by Dennis Yi Tenen
Published 6 Feb 2024

Poetries and Sciences: A Reissue of Science and Poetry (1926, 1935) with Commentary (New York: Norton, 1970), 76–­78. 130 In putting the algorithm in charge: OpenAI, GPT-­4 Technical Report (New York: arXiv, 2023). Index Abelson, Robert, 92 Aeronautical Laboratory (Cornell University), 110 Aesop, 94 agency, 127, 131–32, 141 Agnesi, Maria, 44 AI, See artificial intelligence AI and Data Ethics Institute, 131 Air Research and Development Command (US Air Force), 87 algebraic patterns, 55 algorithms, 9, 131 alignment problem, 38 alphabets, 43 Alphaville (film), 93 American Catalogue of Books in Print, 66 American Psychiatric Association, 23 American Stationer, 74 Analytical Engine (analytical engines), 48–52, 54–56, 60–61, 64 Andrews, Charlton The Technique of the Play Writing, 71 Appelbaum, Matthew, 92 applications, 32, 33 application tables, 48 Arabic language, 43 Arbogast, Louis, 44 Aristotelianism, 34, 36–38, 72 Aristotle, 36, 44 Poetics, 50–51, 67 Ars Brevis (Llull), 24, 31 artifice, 4, 61, 123 artificial intelligence (AI) in academia, 137–38 and agency, 141 as collective labor, 122–23 conversational AI, 135 and creative process, 133–34 dangers of, 127, 129, 137 definitions of, 4, 11 demystification of, 124 economic consequences of, 133–35 ethical, 22 gaps in thinking about, 5–7 history of, 12 “intelligence” aspect of, 14–16, 21, 125 language-based, 21, 46 and machine translation, 119 participants in, 132 personification of, 127, 130 purpose of, 59 and responsibility, 132 artificial intelligence (AI) (continued) scope of, 5, 16, 93, 128, 129 in social sphere, 127, 136, 139 and template culture, 83 assistive technology, 15, 28, 38–39, 123–24, 138 Athenaeum, 74 Austen, Jane, 65, 67 Author, 70 Author’s Digest, 71 Author’s Journal, 70 author wages, 67 automated assistants, 28, 138 Automated Reading of Cursive Script, 110 automated tasks, devaluation of, 38 automatic transmissions, 14–16 automation in industrial age, 2 of reason, 40 of work, 133–34 Babbage, Charles, 43, 48–54, 56, 59–60, 62–64, 71, 105, 118 On the Economy of Machinery and Manufactures, 60, 63–64, 71 Passages from a Life of a Philosopher, 49–50 backstories, 73 Bacon, Francis, 7, 10 Baidu, 113 Baker, George Pierce, 72–73 Baldur’s Gate, 100 Barrymore, John, 73 BASEBALL (story generator), 92 basic drives, 128 Baudot code, 7 Believe Me, Xantippe (film), 73 Bell Telephone Labs, 110 Benjamin, Walter, 61 Bibles, 39 Bibliothèque universelle de Genève, 54 bigrams, 106–7, 109 bits, 6–9 Blackburn, Simon, 84 Bledsoe, W.

…

Our brains do not convert words into numbers or vectors, nor do we have the capacity to iterate over huge numerical datasets. Yet the language models in production today continue to use human-­cognitive metaphors. The machine “learns.” Its statistical connections are called “neurons,” creating a “neural network.” The effects of this technology similarly confound us with human-­like capacity for language, achieved in ways alien to us. I’m therefore suspicious of all metaphors ascribing familiar human cognitive aspects to artificial intelligence. The machine thinks, talks, explains, understands, writes, feels, etc., by analogy only. Its progress holds no explanatory powers for our human ways of being.

…

Their mechanisms of coming to a decision differ entirely from that of humans. In this way, we may speak of “corporate personhood” or a nation being “offended” metaphorically, not in the literal sense of being a person or having feelings. The “intelligence” part of “artificial intelligence” presents a similar condensed figure. Take “machine learning” for example, which Oracle defines as “improving system performance based on consumed data.” Though the metaphor of learning indicates a rough approach, we understand machines to “learn” in ways alien to that of human children. The so-­called “neural” networks producing “learning” effects are mathematical models loosely approximating some aspect of biological brain activity.

What We Owe the Future: A Million-Year View
by William MacAskill
Published 31 Aug 2022

When thinking about lock-in, the key technology is artificial intelligence.35 Writing gave ideas the power to influence society for thousands of years; artificial intelligence could give them influence that lasts millions. I’ll discuss when this might occur later; for now let’s focus on why advanced artificial intelligence would be of such great longterm importance. Artificial General Intelligence Artificial intelligence (AI) is a branch of computer science that aims to design machines that can mimic or replicate human intelligence. Because of the success of machine learning as a paradigm, we’ve made enormous progress in AI over the last ten years.

…

151 Armageddon (film), 106 Arrhenius, Svante, 42 artificial general intelligence (AGI) averting civilisational stagnation, 156 longterm importance of, 80–83 predicting the arrival of, 89–91 prioritising threats to improve on, 228 the pursuit of immortality, 83–86 reducing future uncertainty, 228–229 surpassing human abilities, 86–88 values lock-in, 92–95 artificial intelligence (AI) addressing neglected problems, 231 AI safety, 244 alignment problem, 87 artificial general intelligence, 80–83 defining, 80 future threats and benefits, 6 missing moments of plasticity, 43 prioritising future solutions, 228–229 uncertainty over the future, 224–226 value lock-in, 79 arts and literature preserving and projecting, 22–23 the value of non-wellbeing goods, 214–215 asteroids, collision with, 105–107, 113 Atari, 82–83 Atlantis, 12 Australia: effects of all-out nuclear war, 131 average view of wellbeing, 177–179, 179(fig.)

…

See value lock-in lock-in paradox, 101–102 Long Peace, 114 long reflection, 98–99 longtermism arguments for and against, 4–7, 257–261 concerns for future generations, 10–12 contingency of moral norms, 71–72 empowering future generations, 9 expedition into uncharted terrain, 6–7 longterm consequences of small actions, 173–175 perspective on civilisational stagnation, 159–163 population ethics, 168–171 the size of the future, 19 understanding the implications of, 229–230 values changes, 53–55 lottery winners, 203 Lustig, Richard, 203 lying, negative effects of, 241 Lyons, Oren, 11 Macaulay, Zachary, 69 MacFarquhar, Larissa, 168 machine learning artificial general intelligence development, 80–81 predicting AGI completion, 90–91 See also artificial general intelligence; artificial intelligence mammals evolution of, 4, 13(fig.) lifespan, 13, 13(fig.) megafauna, 29–30 Mao Zedong, 218–219 Marlowe, Frank, 206–207 Mars rovers, 189 mathematics Islamic Golden Age, 143 noncontingency, 32–33 Mauritania: abolition of slavery, 69–70 McKibben, Bill, 43 medicine: expected value theory in decisionmaking, 38 megafauna, 29–30 megatherium, 29 Mercy for Animals, 72–73 Metaculus forecasting platform, 113, 116 metaphors of humanity, history, and longtermism, 6–7 Middle Ages: history of civilisational stagnation, 157 migration.

The Ethical Algorithm: The Science of Socially Aware Algorithm Design
by Michael Kearns and Aaron Roth
Published 3 Oct 2019

Our training notwithstanding, our research and interest in the topics described here have not been formulated in a vacuum of abstraction and mathematics. We’ve both always been interested in applying that approach to problems in machine learning and artificial intelligence. We are also neither adverse to nor inexperienced in experimental, data-driven work in machine learning—often as a test of the practicality and limitations of our theories, but not always. And it was the very trends we describe in these pages—the explosive growth of consumer data enabled by the Internet, and the accompanying rise in machine learning for automated decision-making—that made us and our colleagues aware of and concerned about the potential collateral damage.

…

We don’t need one scientist running a thousand experiments and only misleadingly reporting the results from one of them, because the same thing happens if a thousand scientists each run only one experiment (each in good faith), but only the one with the most surprising result ends up being published. The Sport of Machine Learning The dangers of p-hacking are by no means limited to the traditional sciences: they extend to machine learning as well. Sandy Pentland, a professor at MIT, was quoted in The Economist as saying that “according to some estimates, three-quarters of published scientific papers in the field of machine learning are bunk.” To see a particularly egregious example, let’s go back to 2015, when the market for machine learning talent was heating up. The techniques of deep learning had recently reemerged from relative obscurity (its previous incarnation was called backpropagation in neural networks, which we discussed in the introduction), delivering impressive results in computer vision and image recognition.

…

See lending and creditworthiness crime, 14–15, 62, 92–93 crowdsourcing, 104 cryptography, 31–34, 37 Cuddy, Amy, 141–42 cultural biases, 57 Dalenius, Tore, 35 data administrators, 45–47 data analysis procedures, 164 data collection, 58 dating preferences, 94–97, 100–101 decision-making process, 3–4, 11, 190–91 decision trees, 154–55, 159–60, 164, 173 deep learning algorithms, 133, 146–47 DeepMind, 179 deep neural networks, 174–75 defection, 99–100, 115, 128 demographics, 7, 193 derived models, 6 diagnostics, 27–29 dietary research, 143–45, 158 differential privacy to combat overfitting data, 167 commercial deployment of, 47–49 and correlation equilibrium, 114–15 described, 36–39 design of ethical algorithms, 193, 195 and differing notions of fairness, 85 and embarrassing polls, 40, 43–45 and fairness vs. accuracy of models, 63 and game-theoretic algorithm design, 135 limitations of, 50–56 and trust in data administrators, 45–47 Dijkstra’s algorithm, 104, 109 diminishing marginal returns, 186–88 disability status, 86–89 discrimination and algorithmic violations of fairness and privacy, 96 and data collection bias, 90–93 and “fairness gerrymandering,” 86–89 and fairness issues in machine learning, 65–66 and fairness vs. accuracy of models, 63 and game-theoretic algorithm design, 134–35 and “merit” in algorithmic fairness, 75 and recent efforts to address machine learning issues, 15 and self-play in machine learning, 132–33 and statistical parity, 69 and supervised machine learning, 64 and unique challenges of algorithms, 7 and user preferences, 96–97. See also gender data and bias; racial data and bias disease prediction models, 30–33, 54, 96 diversity, 125–26 divorce proceedings, 40–41, 48, 51–52 DNA data, 30–31, 54–56 Doll, Richard, 34 doomsday scenarios, 184–85 downstream effects, 194 driving times.

AI Superpowers: China, Silicon Valley, and the New World Order
by Kai-Fu Lee
Published 14 Sep 2018

In that connection—human beings giving and receiving love—I caught a glimpse of how humans will find work and meaning in the age of artificial intelligence. I believe that the skillful application of AI will be China’s greatest opportunity to catch up with—and possibly surpass—the United States. But more important, this shift will create an opportunity for all people to rediscover what it is that makes us human. To understand why, we must first grasp the basics of the technology and how it is set to transform our world. A BRIEF HISTORY OF DEEP LEARNING Machine learning—the umbrella term for the field that includes deep learning—is a history-altering technology but one that is lucky to have survived a tumultuous half-century of research.

…

In that touchscreen device and that unmet desire for human contact, I saw the first sketches of a blueprint for coexistence between people and artificial intelligence. Yes, intelligent machines will increasingly be able to do our jobs and meet our material needs, disrupting industries and displacing workers in the process. But there remains one thing that only human beings are able to create and share with one another: love. With all of the advances in machine learning, the truth remains that we are still nowhere near creating AI machines that feel any emotions at all. Can you imagine the elation that comes from beating a world champion at the game you’ve devoted your whole life to mastering?

…

v=UF8uR6Z6KLc&t=785s. space race of the 1960s: John R. Allen and Amir Husain, “The Next Space Race Is Artificial Intelligence: And the United States Is Losing,” Foreign Policy, November 3, 2017, http://foreignpolicy.com/2017/11/03/the-next-space-race-is-artificial-intelligence-and-america-is-losing-to-china/. Cold War arms race: Zachary Cohen, “US Risks Losing Artificial Intelligence Arms Race to China and Russia,” CNN, November 29, 2017, https://www.cnn.com/2017/11/29/politics/us-military-artificial-intelligence-russia-china/index.html. Index A | B | C | D | E | F | G | H | I | J | K | L | M | N | O | P | Q | R | S | T | U | V | W | X | Y | Z A Africa, 138, 139, 169 age of data, 14, 18, 56 age of implementation Chinese entrepreneurs and, 16, 18, 25 Chinese government and, 18 data and, 17, 20, 55, 80 deep learning and, 13–14, 143 going light vs. going heavy, 71 AGI (artificial general intelligence), 140–44 AI.

Machine Translation
by Thierry Poibeau
Published 14 Sep 2017

Facebook bought out different companies specialized in machine translation (such as Jibbigo in 2013 for voice messages in particular). Apple and Google are also regularly buying startups in the communication and information technology domains. Most importantly, all these large companies are hiring engineers and researchers (mainly in machine learning and artificial intelligence) in order to produce their own machine translation solution. They are also opening new research centers worldwide in order to attract the best talent everywhere. New Applications of Machine Translation The machine translation market is growing fast. Over the last few years we have witnessed the emergence of new applications, particularly on mobile devices.

…

Since the advent of computers (after the Second World War), this research program has materialized through the design of machine translation tools—in other words, computer programs capable of automatically producing in a target language the translation of a text in a source language. This research program is very ambitious: it is even one of the most fundamental in the field of artificial intelligence. The analysis of languages cannot be separated from the analysis of knowledge and reasoning, which explains the interest in this field shown by philosophers and specialists of artificial intelligence as well as the cognitive sciences. This brings to mind the test proposed by Turing3 in 1950: the test is successfully completed if a person dialoguing (through a screen) with a computer is unable to say whether her discussion partner is a computer or a human being.

…

In fact, it is possible to generalize the approach so to consider the problem of translation as an alignment problem at the level of sequences of words, and not at the level of isolated words only. The goal is to translate at the phrase level (i.e., sequences of several words): this would enable the context to be better taken into account and would thus offer translations of better quality than simple word-for-word equivalences. It is possible to generalize the approach so as to consider the problem of translation as an alignment problem at the level of sequences of words, and not at the level of isolated words only.

The Singularity Is Nearer: When We Merge with AI
by Ray Kurzweil
Published 25 Jun 2024

,” MIS Quarterly 19, no. 1 (March 1995): 51–81, https://www.jstor.org/stable/249711. BACK TO NOTE REFERENCE 21 For a short and nontechnical explainer on why machine learning reduces the complexity ceiling problem, see Deepanker Saxena, “Machine Learning vs. Rules Based Systems,” Socure, August 6, 2018, https://www.socure.com/blog/machine-learning-vs-rule-based-systems. BACK TO NOTE REFERENCE 22 Cade Metz, “One Genius’ Lonely Crusade to Teach a Computer Common Sense,” Wired, March 24, 2016, https://www.wired.com/2016/03/doug-lenat-artificial-intelligence-common-sense-engine; “Frequently Asked Questions,” Cycorp, accessed November 20, 2021, https://cyc.com/faq.

…

BACK TO NOTE REFERENCE 171 For several more detailed summaries of recent applications of AI to materials discovery for solar electricity and energy storage, see Elizabeth Montalbano, “AI Enables Design of Spray-on Coating That Can Generate Solar Energy,” Design News, December 26, 2019, https://www.designnews.com/materials-assembly/ai-enables-design-spray-coating-can-generate-solar-energy; Shinji Nagasawa, Eman Al-Naamani, and Akinori Saeki, “Computer-Aided Screening of Conjugated Polymers for Organic Solar Cell: Classification by Random Forest,” Journal of Physical Chemistry Letters 9, no. 10 (May 7, 2018): 2639–46, https://doi.org/10.1021/acs.jpclett.8b00635; Geun Ho Gu et al., “Machine Learning for Renewable Energy Materials,” Journal of Materials Chemistry A 7, no. 29 (April 30, 2019): 17096–117, https://doi.org/10.1039/C9TA02356A; Ziyi Luo et al., “A Survey of Artificial Intelligence Techniques Applied in Energy Storage Materials R&D,” Frontiers in Energy Research 8, no. 116 (July 3, 2020), https://doi.org/10.3389/fenrg.2020.00116; An Chen, Xu Zhang, and Zhen Zhou, “Machine Learning: Accelerating Materials Development for Energy Storage and Conversion,” InfoMat 2, no. 3 (February 23, 2020): 553–76, https://doi.org/10.1002/inf2.12094; Xinyi Yang et al., “Development Status and Prospects of Artificial Intelligence in the Field of Energy Conversion Materials,” Frontiers in Energy Research 8, no. 167 (July 31, 2020), https://doi.org/10.3389/fenrg.2020.00167; Teng Zhou, Zhen Song, and Kai Sundmacher, “Big Data Creates New Opportunities for Materials Research: A Review on Methods and Applications of Machine Learning for Materials Design,” Engineering 5, no. 6 (December 2019): 1017–26, https://doi.org/10.1016/j.eng.2019.02.011.

…

v=JqYte9UMJCg; Pranav Rajpurkar et al., “CheXNet: Radiologist-Level Pneumonia Detection on Chest X-Rays with Deep Learning,” Stanford Machine Learning Group working paper, November 14, 2017, https://arxiv.org/pdf/1711.05225v1.pdf; Jeremy Irvin et al., “CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparison,” Proceedings of the AAAI Conference on Artificial Intelligence 33, no. 1 (July 17, 2019): AAAI-10, IAAI-19, EAAI-20, https://www.aaai.org/ojs/index.php/AAAI/article/view/3834. BACK TO NOTE REFERENCE 35 Huiying Liang et al., “Evaluation and Accurate Diagnoses of Pediatric Diseases Using Artificial Intelligence,” Nature Medicine 25, no. 3 (February 11, 2019): 433–38, https://doi.org/10.1038/s41591-018-0335-9.

Rule of the Robots: How Artificial Intelligence Will Transform Everything
by Martin Ford
Published 13 Sep 2021

Watson heralded a new age and portended machines that would finally begin to parse language and truly engage with humans, but 2011 would also mark the beginning of a dramatic shift in the underlying technology of artificial intelligence. Watson relied on machine learning algorithms that used statistical techniques to make sense of information, but over the next few years, another kind of machine learning—based directly on the perceptron conceived by Frank Rosenblatt more than half a century earlier—would once again come to the forefront and then rapidly rise to dominate the field of artificial intelligence. CONNECTIONIST VS. SYMBOLIC AI AND THE RISE OF DEEP LEARNING Even as the general field of artificial intelligence traced its boom-and-bust path over the decades, the research focus swung between two general philosophies that emphasized contrasting approaches to building more intelligent machines.

…

In other words, this preliminary version of AlphaFold was not yet accurate enough to be a truly useful research tool.80 The fact that DeepMind was able to refine its technology to the point where a number of scientists declared the protein folding problem to be “solved” just two years later is, I think, an especially vivid indication of just how rapidly specific applications of artificial intelligence are likely to continue advancing. Aside from using machine learning to discover new drugs and other chemical compounds, the most promising general application of artificial intelligence to scientific research may be in the assimilation and understanding of the continuously exploding volume of published research. In 2018 alone, more than three million scientific papers were published in more than 40,000 separate journals.81 Making sense of information on that scale is so far beyond the capability of any individual human mind that artificial intelligence is arguably the only tool at our disposal that could lead to some sort of holistic comprehension.

…

BIAS, FAIRNESS AND TRANSPARENCY IN MACHINE LEARNING ALGORITHMS As artificial intelligence and machine learning are deployed more and more widely, it’s critical that the results and recommendations produced by these algorithms are perceived as fair and that the reasoning behind them can be adequately explained. If you’re using a deep learning system to maximize the energy efficiency of some industrial machine, then you are probably not particularly concerned about the details that drive an algorithmic outcome; you simply want the optimal result. But when machine learning is applied to areas like criminal justice, hiring decisions or the processing of home mortgage applications—in other words, to high-stakes decisions that directly impact the rights and future well-being of human beings—it’s essential that algorithmic outcomes can be shown to be unbiased across demographic groups and that the analysis that led to those outcomes is transparent and just.

Artificial Intelligence: A Guide for Thinking Humans
by Melanie Mitchell
Published 14 Oct 2019

A recent AI survey paper summed it up: “Because we don’t deeply understand intelligence or know how to produce general AI, rather than cutting off any avenues of exploration, to truly make progress we should embrace AI’s ‘anarchy of methods.’”14 But since the 2010s, one family of AI methods—collectively called deep learning (or deep neural networks)—has risen above the anarchy to become the dominant AI paradigm. In fact, in much of the popular media, the term artificial intelligence itself has come to mean “deep learning.” This is an unfortunate inaccuracy, and I need to clarify the distinction. AI is a field that includes a broad set of approaches, with the goal of creating machines with intelligence. Deep learning is only one such approach. Deep learning is itself one method among many in the field of machine learning, a subfield of AI in which machines “learn” from data or from their own “experiences.” To better understand these various distinctions, it’s important to understand a philosophical split that occurred early in the AI research community: the split between so-called symbolic and subsymbolic AI.

…

The original goals of AI—computers that could converse with us in natural language, describe what they saw through their camera eyes, learn new concepts after seeing only a few examples—are things that young children can easily do, but, surprisingly, these “easy things” have turned out to be harder for AI to achieve than diagnosing complex diseases, beating human champions at chess and Go, and solving complex algebraic problems. As Minsky went on, “In general, we’re least aware of what our minds do best.”27 The attempt to create artificial intelligence has, at the very least, helped elucidate how complex and subtle are our own minds. 2 Neural Networks and the Ascent of Machine Learning Spoiler alert: Multilayer neural networks—the extension of perceptrons that was dismissed by Minsky and Papert as likely to be “sterile”—have instead turned out to form the foundation of much of modern artificial intelligence. Because they are the basis of several of the methods I’ll describe in later chapters, I’ll take some time here to describe how these networks work.

…

Nagy, “Neural Networks—Then and Now,” IEEE Transactions on Neural Networks 2, no. 2 (1991): 316–18. 24.  Minsky and Papert, Perceptrons, 231–32. 25.  J. Lighthill, “Artificial Intelligence: A General Survey,” in Artificial Intelligence: A Paper Symposium (London: Science Research Council, 1973). 26.  Quoted in C. Moewes and A. Nürnberger, Computational Intelligence in Intelligent Data Analysis (New York: Springer, 2013), 135. 27.  M. L. Minsky, The Society of Mind (New York: Simon & Schuster, 1987), 29. 2: Neural Networks and the Ascent of Machine Learning   1.  The activation value y at each hidden and output unit is typically computed by taking the dot product between the vector x of inputs to the unit and the vector w of weights on the connections to that unit, and applying the sigmoid function to the result: y = 1/(1 + e−(x.w)).

A Hacker's Mind: How the Powerful Bend Society's Rules, and How to Bend Them Back
by Bruce Schneier
Published 7 Feb 2023

The trick is figuring out how to encourage the good hacks while stopping the bad ones, and knowing the difference between the two. Hacking will become even more disruptive as we increasingly implement artificial intelligence (AI) and autonomous systems. These are computer systems, which means they will inevitably be hacked in the same ways that all computer systems are. They affect social systems—already AI systems make loan, hiring, and parole decisions—which means those hacks will consequently affect our economic and political systems. More significantly, machine-learning processes that underpin all of modern AI will result in the computers performing the hacks. Extrapolating further, AI systems will soon start discovering new hacks.

…

Luke Halpin and Doug Dannemiller (2019), “Artificial intelligence: The next frontier for investment management firms,” Deloitte, https://www2.deloitte.com/content/dam/Deloitte/global/Documents/Financial-Services/fsi-artificial-intelligence-investment-mgmt.pdf. Peter Salvage (March 2019), “Artificial intelligence sweeps hedge funds,” BNY Mellon, https://www.bnymellon.com/us/en/insights/all-insights/artificial-intelligence-sweeps-hedge-funds.html. 244the precautionary principle: Maciej Kuziemski (1 May 2018), “A precautionary approach to artificial intelligence,” Project Syndicate, https://www.project-syndicate.org/commentary/precautionary-principle-for-artificial-intelligence-by-maciej-kuziemski-2018-05. 60.

…

Public services, business transactions, and even basic social interactions are now mediated by digital systems that make predictions and decisions just like humans do, but they do it faster, more consistently, and less accountably than humans. Our machines increasingly make decisions for us, but they don’t think like we do, and the interaction of our minds with these artificial intelligences points the way to an exciting and dangerous future for hacking: in the economy, the law, and beyond. PART 7 HACKING AI SYSTEMS 50 Artificial Intelligence and Robotics Artificial intelligence—AI—is an information technology. It consists of software, it runs on computers, and it is already deeply embedded into our social fabric, both in ways we understand and in ways we don’t.

12 Bytes: How We Got Here. Where We Might Go Next
by Jeanette Winterson
Published 15 Mar 2021

Wells, 1898 People, Power, and Profits: Progressive Capitalism for an Age of Discontent, Joseph Stiglitz, 2019 The Sixth Extinction: An Unnatural History, Elizabeth Kolbert, 2014 Utopia for Realists: The Case for a Universal Basic Income, Open Borders, and a 15-hour Workweek, 2014, and Humankind: A Hopeful History, 2019, Rutger Bregman Notes from an Apocalypse: A Personal Journey to the End of the World and Back, Mark O’Connell, 2020 The Better Angels of Our Nature: Why Violence Has Declined, Steven Pinker, 2011 Blockchain Chicken Farm: And Other Stories of Tech in China’s Countryside, Xiaowei Wang, 2020 Life 3.0: Being Human in the Age of Artificial Intelligence, Max Tegmark, 2017 The Alignment Problem: How Can Machines Learn Human Values?, Brian Christian, 2021 I Love, Therefore I Am There is no reading list. There is everything you are. Illustration and Text Credits p.13 Babbage’s Difference Engine No 1, 1824–1832 © Science & Society Picture Library / Getty Images; p.14 Punch cards on a Jacquard loom © John R.

…

The smarter AI gets, the smarter robots will get. At present there are serious technical issues to overcome. All artificial intelligence is narrow AI – programmed, specific, problem-solving that doesn’t transfer well to other domains. It’s not AGI, where the system would operate more like a human brain does. A robot with a map of your kitchen wouldn’t ‘know’ why a table is where it is – and it will be confused if the table moves. Statistical knowledge is not the same thing as general understanding. Machine learning can get around this by throwing more data at the problem (train your AI on bigger data-sets), but we aren’t solving the underlying issue.

…

* * * Bloom points out that most humans are fixated on space without boundaries. Think about it: land-grab, colonisation, urban creep, loss of habitat, the current fad for seasteading (sea cities with vast oceans at their disposal). And space itself – the go-to fascination of rich men: Richard Branson, Elon Musk, Jeff Bezos. When I think about artificial intelligence, and what is surely to follow – artificial general intelligence, or superintelligence – it seems to me that what this affects most, now and later, isn’t space but time. The brain uses chemicals to transmit information. A computer uses electricity. Signals travel at high speeds through the nervous system (neurons fire 200 times a second, or 200 hertz) but computer processors are measured in gigahertz – billions of cycles per second.

The Coming Wave: Technology, Power, and the Twenty-First Century's Greatest Dilemma
by Mustafa Suleyman
Published 4 Sep 2023

GO TO NOTE REFERENCE IN TEXT “By 2030, China’s AI theories” Graham Webster et al., “Full Translation: China’s ‘New Generation Artificial Intelligence Development Plan,’ ” DigiChina, Stanford University, Aug. 1, 2017, digichina.stanford.edu/​work/​full-translation-chinas-new-generation-artificial-intelligence-development-plan-2017. GO TO NOTE REFERENCE IN TEXT Indeed, Tsinghua publishes more Benaich and Hogarth, State of AI; Neil Savage, “The Race to the Top Among the World’s Leaders in Artificial Intelligence,” Nature Index, Dec. 9, 2020, www.nature.com/​articles/​d41586-020-03409-8; “Tsinghua University May Soon Top the World League in Science Research,” Economist, Nov. 17, 2018, www.economist.com/​china/​2018/​11/​17/​tsinghua-university-may-soon-top-the-world-league-in-science-research.

…

See Laura Cooper and Preeti Singh, “Private Equity Backs Record Volume of Tech Deals,” Wall Street Journal, Jan. 3, 2022, www.wsj.com/​articles/​private-equity-backs-record-volume-of-tech-deals-11641207603. GO TO NOTE REFERENCE IN TEXT Investment in AI technologies See, for example, Artificial Intelligence Index Report 2021, although the numbers have certainly grown in the generative AI boom since then. GO TO NOTE REFERENCE IN TEXT PwC forecasts AI will add “Sizing the Prize—PwC’s Global Artificial Intelligence Study: Exploiting the AI Revolution,” PwC, 2017, www.pwc.com/​gx/​en/​issues/​data-and-analytics/​publications/​artificial-intelligence-study.html. GO TO NOTE REFERENCE IN TEXT McKinsey forecasts a $4 trillion boost Jacques Bughin et al., “Notes from the AI Frontier: Modeling the Impact of AI on the World Economy,” McKinsey, Sept. 4, 2018, www.mckinsey.com/​featured-insights/​artificial-intelligence/​notes-from-the-ai-frontier-modeling-the-impact-of-ai-on-the-world-economy; Michael Ciu, “The Bio Revolution: Innovations Transforming Economies, Societies, and Our Lives,” McKinsey Global Institute, May 13, 2020, www.mckinsey.com/​industries/​pharmaceuticals-and-medical-products/​our-insights/​the-bio-revolution-innovations-transforming-economies-societies-and-our-lives.

…

GO TO NOTE REFERENCE IN TEXT The most ambitious legislation “The Artificial Intelligence Act,” Future of Life Institute, artificialintelligenceact.eu. GO TO NOTE REFERENCE IN TEXT Some argue it’s too focused See, for example, “FLI Position Paper on the EU AI Act,” Future of Life Institute, Aug. 4, 2021, futureoflife.org/​wp-content/​uploads/​2021/​08/​FLI-Position-Paper-on-the-EU-AI-Act.pdf?x72900; and David Matthews, “EU Artificial Intelligence Act Not ‘Futureproof,’ Experts Warn MEPs,” Science Business, March 22, 2022, sciencebusiness.net/​news/​eu-artificial-intelligence-act-not-futureproof-experts-warn-meps.

Enlightenment Now: The Case for Reason, Science, Humanism, and Progress
by Steven Pinker
Published 13 Feb 2018

See also note 25 above. 27. Shallowness and brittleness of current AI: Brooks 2015; Davis & Marcus 2015; Lanier 2014; Marcus 2016; Schank 2015. 28. Naam 2010. 29. Robots turning us into paper clips and other Value Alignment Problems: Bostrom 2016; Hanson & Yudkowsky 2008; Omohundro 2008; Yudkowsky 2008; P. Torres, “Fear Our New Robot Overlords: This Is Why You Need to Take Artificial Intelligence Seriously,” Salon, May 14, 2016. 30. Why we won’t be turned into paper clips: B. Hibbard, “Reply to AI Risk,” http://www.ssec.wisc.edu/~billh/g/AIRisk_Reply.html; R. Loosemore, “The Maverick Nanny with a Dopamine Drip: Debunking Fallacies in the Theory of AI Motivation,” Institute for Ethics and Emerging Technologies, July 24, 2014, http://ieet.org/index.php/IEET/more/loosemore20140724; A.

…

See also Holocaust Anton, Michael (“Publius Decius Mus”), 448, 449 anxiety, 283 adulthood and, 288–9 “collapse anxiety,” 292 depression as comorbid with, 283 and institutions, loss of faith in, 286 media practices of encouraging, 287 as motivation to solve problems, 287 postwar increase in, 284 prevalence of depression and, 282–3, 476n74 sex differences in, 285 strategems for coping with, 287 women’s gains in autonomy and, 285 Appiah, Kwame Anthony, 443 Aquino, Corazon, 91 Arab countries classical Arab civilization, 439, 442 clerical meddling in education, 234 slavery/racism and, 397 See also Muslim countries Arab Spring (2011), 203, 228, 370 archaeology, 407 Argentina, 200, 315 Ariely, Dan, 353 Aristotle, eudaemonia, 267 Arkhipov, Vasili, 479n93 Armenia, 158 artificial intelligence (AI) “Artificial General Intelligence” (AGI), 297, 298 and Enlightenment thinkers, 386 as existential threat, putative, 296–300, 477n20 job losses and, 118, 300 Value Alignment Problem, 299–300 arts and culture availability of, 260–61 and consilience with science, 407–9 depicting traditionalism vs. modernity, 284 ideological innumeracy and, 48 Nietzsche as influence on, 445, 446–7 vs. science, 34, 389–90 Aryans, romantic heroism and, 33, 398, 444 Asafu-Adjaye, John, 122 Asia authoritarian regimes, rise of, 200 carbon emissions of, 144 famine in, 69, 78 globalization and, 111, 112, 117 IQ gains in, 241 life expectancy in, 53–4, 54, 55 military governments of, 200 postcolonial governments of, 78 undernourishment in, 72 See also individual countries and subregions Asians, hate crimes against, 219, 220 Asiri, Abdullah al-, 303 Assad, Bashar al-, 159 Astell, Mary, 252 atheism and atheists charitable acts by, 432 dangers of self-labeling as, 435 definition of, 430 moral realism of, 429 “New Atheism,” 430 numbers of, 435, 436, 437–8, 489n68, 490n65 rising Intelligence Quotient test scores and, 438 wars by, 429–30 Athens (ancient), 212 Atkins, Peter, 17 Auden, W.

…

This is the danger that we will be subjugated, intentionally or accidentally, by artificial intelligence (AI), a disaster sometimes called the Robopocalypse and commonly illustrated with stills from the Terminator movies. As with Y2K, some smart people take it seriously. Elon Musk, whose company makes artificially intelligent self-driving cars, called the technology “more dangerous than nukes.” Stephen Hawking, speaking through his artificially intelligent synthesizer, warned that it could “spell the end of the human race.”19 But among the smart people who aren’t losing sleep are most experts in artificial intelligence and most experts in human intelligence.20 The Robopocalypse is based on a muzzy conception of intelligence that owes more to the Great Chain of Being and a Nietzschean will to power than to a modern scientific understanding.21 In this conception, intelligence is an all-powerful, wish-granting potion that agents possess in different amounts.

Artificial Intelligence: A Modern Approach
by Stuart Russell and Peter Norvig
Published 14 Jul 2019

The most recent work appears in the proceedings of the major AI conferences: the International Joint Conference on AI (IJCAI), the annual European Conference on AI (ECAI), and the AAAI Conference. Machine learning is covered by the International Conference on Machine Learning and the Neural Information Processing Systems (NeurIPS) meeting. The major journals for general AI are Artificial Intelligence, Computational Intelligence, the IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Intelligent Systems, and the Journal of Artificial Intelligence Research. There are also many conferences and journals devoted to specific areas, which we cover in the appropriate chapters. 1In the public eye, there is sometimes confusion between the terms “artificial intelligence” and “machine learning.”

…

Bibliographical and Historical Notes A comprehensive history of AI is given by Nils Nilsson (2009), one of the early pioneers of the field. Pedro Domingos (2015) and Melanie Mitchell (2019) give overviews of machine learning for a general audience, and Kai-Fu Lee (2018) describes the race for international leadership in AI. Martin Ford (2018) interviews 23 leading AI researchers. The main professional societies for AI are the Association for the Advancement of Artificial Intelligence (AAAI), the ACM Special Interest Group in Artificial Intelligence (SIGAI, formerly SIGART), the European Association for AI, and the Society for Artificial Intelligence and Simulation of Behaviour (AISB). The Partnership on AI brings together many commercial and nonprofit organizations concerned with the ethical and social impacts of AI.

…

Learning dependency–based compositional semantics. arXiv:1109.6841. Liang, P. and Potts, C. (2015). Bringing machine learning and compositional semantics together. Annual Review of Linguistics, 1, 355–376. Liberatore, P. (1997). The complexity of the language A. Electronic Transactions on Artificial Intelligence, 1, 13–38. Lifschitz, V. (2001). Answer set programming and plan generation. AIJ, 138, 39–54. Lighthill, J. (1973). Artificial intelligence: A general survey. In Lighthill, J., Sutherland, N. S., Needham, R. M., Longuet–Higgins, H. C., and Michie, D. (Eds.), Artificial Intelligence: A Paper Symposium. Science Research Council of Great Britain.

The Precipice: Existential Risk and the Future of Humanity
by Toby Ord
Published 24 Mar 2020

The most plausible existential risk would come from success in AI researchers’ grand ambition of creating agents with a general intelligence that surpasses our own. But how likely is that to happen, and when? In 2016, a detailed survey was conducted of more than 300 top researchers in machine learning.84 Asked when an AI system would be “able to accomplish every task better and more cheaply than human workers,” on average they estimated a 50 percent chance of this happening by 2061 and a 10 percent chance of it happening as soon as 2025.85 FIGURE 5.1 Measures of progress and interest in artificial intelligence. The faces show the very rapid recent progress in generating realistic images of “imagined” people. The charts show longterm progress in chess AI surpassing the best human grand masters (measured in Elo), as well as the recent rise in academic activity in the field—measured by papers posted on arXiv, and attendance at conferences.86 This should be interpreted with care.

…

The international body responsible for the continued prohibition of bioweapons (the Biological Weapons Convention) has an annual budget of just $1.4 million—less than the average McDonald’s restaurant.54 The entire spending on reducing existential risks from advanced artificial intelligence is in the tens of millions of dollars, compared with the billions spent on improving artificial intelligence capabilities.55 While it is difficult to precisely measure global spending on existential risk, we can state with confidence that humanity spends more on ice cream every year than on ensuring that the technologies we develop do not destroy us.56 In scientific research, the story is similar.

…

While exploring society’s current vulnerabilities or the dangers from recent techniques, the biosecurity community also emits dangerous information (something I’ve had to be acutely aware of while writing this section).70 This makes the job of those trying to protect us even harder. UNALIGNED ARTIFICIAL INTELLIGENCE In the summer of 1956 a small group of mathematicians and computer scientists gathered at Dartmouth College to embark on the grand project of designing intelligent machines. They explored many aspects of cognition including reasoning, creativity, language, decision-making and learning. Their questions and stances would come to shape the nascent field of artificial intelligence (AI). The ultimate goal, as they saw it, was to build machines rivaling humans in their intelligence.71 As the decades passed and AI became an established field, it lowered its sights.

Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity
by Daron Acemoglu and Simon Johnson
Published 15 May 2023

“Zero-Hours Contracts Have a Devastating Impact on Career Progression—Labour Is Right to Ban Them.” Conversation, September 24. https://theconversation.com/zero-hours-contracts-have-a-devastating-impact-on-career-progression-labour-is-right-to-ban-them-123066. Neapolitan, Richard E., and Xia Jiang. 2018. Artificial Intelligence: With an Introduction to Machine Learning, 2nd ed. London: Chapman and Hall/CRC. Neeson, J. M. 1993. Commoners, Common Right, Enclosure and Social Change in England, 1700‒1820. Cambridge: Cambridge University Press. Noble, David. 1977. America by Design: Science, Technology, and the Rise of Corporate Capitalism.

…

Medium, November 27, https://medium.com/@francois.chollet/the-impossibility-of-intelligence-explosion-5be4a9eda6ec. Chollet, François. 2019. “On the Measure of Intelligence.” Working paper, https://arxiv.org/pdf/1911.01547.pdf?ref=https://githubhelp.com. Christian, Brian. 2020. The Alignment Problem: Machine Learning and Human Values. New York: W.W. Norton. Chudek, Maciej, Sarah Heller, Susan Birch, and Joseph Henrich. 2012. “Prestige-Biased Cultural Learning: Bystander’s Differential Attention to Potential Models Influences Children’s Learning.” Evolution and Human Behavior 33, no. 1: 46‒56. Cialdini, Robert B. 2006.

…

If a company is installing, say, new computers, this must mean that the higher revenues they generate more than make up for the costs. But in a world in which shared visions guide our actions, there is no guarantee that this is indeed the case. If everybody becomes convinced that artificial-intelligence technologies are needed, then businesses will invest in artificial intelligence, even when there are alternative ways of organizing production that could be more beneficial. Similarly, if most researchers are working on a particular way of advancing machine intelligence, others may follow faithfully, or even blindly, in their footsteps.

Practical C Programming, 3rd Edition
by Steve Oualline
Published 15 Nov 2011

Although trees are supposed to be fast, this program was so slow that you would think I used a linked list. Why? Hint: Graphically construct a tree using the words “able,” “baker,” “cook,” “delta,” and “easy,” and look at the result. (Click here for the answer Section 17.11) Data Structures for a Chess Program One of the classic problems in artificial intelligence is the game of chess. As this book goes to press, the Grandmaster who beat the world’s best chess-playing computer last year has lost to the computer this year (1997). We are going to design a data structure for a chess-playing program. In chess, you have several possible moves that you can make.

…

The program can automatically fix the file problem: const long int MAGIC = 0x11223344L /* file identification number*/ const long int SWAP_MAGIC = 0x22114433L /* magic-number byte swapped */ FILE *in_file; /* file containing binary data */ long int magic; /* magic number from file */ in_file = fopen("data", "rb"); fread((char *)&magic, sizeof(magic), 1, in_file); switch (magic) { case MAGIC: /* No problem */ break; case SWAP_MAGIC: printf("Converting file, please wait\n"); convert_file(in_file); break; default: fprintf(stderr,"Error:Bad magic number %lx\n", magic); exit (8); } Alignment Problem Some computers limit the address that can be used for integers and other types of data. For example, the 68000 series require that all integers start on a 2-byte boundary. If you attempt to access an integer using an odd address, you will generate an error. Some processors have no alignment rules, while some are even more restrictive—requiring integers to be aligned on a 4-byte boundary.

…

. ==, How to Learn C == operator (equal), if Statement, Debugging, General > operator (greater than), if Statement >= operator (greater than or equal to), if Statement \\>\\> (right shift), Bit Operators, The Left- and Right-Shift Operators (<<, >>) \\n (newline character), Characters ^ (exclusive or), Bit Operators, The Bitwise Exclusive or (^) __MSDOS__ (pre-defined symbol), Conditional Compilation __STDC_ (pre-defined symbol), Conditional Compilation __TURBOC_ (pre-defined symbol), Conditional Compilation {} (curly braces), if Statement | (or operator), Bitwise or (|) | (or), Bit Operators ~ (complement operator), The Ones Complement Operator (Not) (~) ~ (complement), Bit Operators A accuracy, floating point, Accuracy, Determining Accuracy addition operator (+), Simple Expressions addition, floating point, Floating Addition/Subtraction address of operator (&), Simple Pointers address operator (&), Simple Pointers alignment restrictions, Alignment Problem ambiguous code, else Statement and operator (&), Bit Operators, The and Operator (&) and portability, Byte Order Problem Appendix A, ASCII Table, Characters argc, Command-Line Arguments argv, Command-Line Arguments array, declarations, How C Works arrays, Arrays, Arrays of Structures and pointers, Pointers and Arrays index, Arrays infinite, A Program to Use Infinite Arrays, Using the Infinite Array initializing, Initializing Variables multiple dimensional, Initializing Variables of structures, Arrays of Structures arrays, dimension, Arrays arrays, element, Arrays ASCII characters, Characters files, Binary and ASCII Files versus binary files, Binary and ASCII Files assembly language, How Programming Works assembly language translation, How Programming Works assignment statement, How C Works assignment statements, General author, Style auto, Scope and Class automatic parameter changes, Pointers and Arrays variables, Scope and Class B binary files, Binary and ASCII Files, Byte Order Problem I/O, Binary I/O mode for fopen (b), The End-of-Line Puzzle trees, Trees bit, Bit Operations bit fields, Bit Fields or Packed Structures bit operations, Bit Operations bit operators, Bit Operators Bit values, Setting, Clearing, and Testing Bits bitmapped graphics, Bitmapped Graphics bitmaps, Bitmapped Graphics bits, Integers bits, clearing, Setting, Clearing, and Testing Bits bits, defining, Setting, Clearing, and Testing Bits bits, setting, Setting, Clearing, and Testing Bits bits, testing, Setting, Clearing, and Testing Bits bitwise and (&), Bit Operators, The and Operator (&) bitwise compement (~), Bit Operators bitwise complement (~), The Ones Complement Operator (Not) (~) bitwise exclusive or (^), Bit Operators, The Bitwise Exclusive or (^) bitwise left shift (), Bit Operators bitwise or (|), Bit Operators, Bitwise or (|) bitwise right shift (\\>\\>), Bit Operators, The Left- and Right-Shift Operators (<<, >>) bitwise shift left (), The Left- and Right-Shift Operators (<<, >>) blank modifier, The extern Modifier boolean algebra, Bit Operations bottom-up programming, Structured Programming Bourne shell, #define Statement branching statements, Decision and Control Statements break statement, break Statement, switch Statement breakpoints, Debugging a Binary Search buffered file problems, Buffering Problems byte, Bit Operations bytes, Byte Order Problem C C language, Brief History of C C tools, Electronic Archaeology C++ language, Brief History of C C++, how to learn, How to Learn C Calc (program), Specification, Prototype calc (programmed with switch statements), switch Statement calculation operators, Simple Expressions call graphs, Electronic Archaeology carriage return, The End-of-Line Puzzle case labels, switch Statement case statement, switch Statement PASCAL, switch Statement cb program, Electronic Archaeology cc command, #define Statement, Conditional Compilation cc program, C Preprocessor cd command, Setting Up cdb debugger, Interactive Debuggers cflow program, Electronic Archaeology char, Characters (as integer), Types of Integers character constants, Characters characters, Characters chess, data structures for, Data Structures for a Chess Program class, Scope and Class class, derrived, Line Counter Submodule (lc) class, variable, Scope and Class classes, How This Book is Organized clause conditional, Decision and Control Statements clearing bits, Setting, Clearing, and Testing Bits close, Unbuffered I/O COBOL, How Programming Works code commenting out, Conditional Compilation design, Programming Process, Code Design format, Indentation and Code Format, Simplicity maintaining, Style code design, Specification coding, Coding command-line arguments, Command-Line Arguments command-line arguments, sample program, Command-Line Arguments commands, Interactive Debuggers dbx, Interactive Debuggers commenting out code, Conditional Compilation comments, How to Learn C, Style boxes, Style program, Add Comments comments, author, Style comments, file formats, Style comments, functions, Functions comments, heading, Style, Basic Program Structure comments, in data files, Designing File Formats comments, notes, Style comments, procedures, Functions comments, purpose, Style comments, references, Style comments, restrictions, Style comments, revision history, Style comments, units, Common Coding Practices compile command, Conditional Compilation compiler, How C Works compilers, Compiler conditional, Conditional Compilation complement (~), Bit Operators, The Ones Complement Operator (Not) (~) complex data types, How C Works computation operators, Simple Expressions concatenating strings, Strings conditional clause, Decision and Control Statements conditionals, General const, Constant Declarations const pointers, const Pointers const vs.

Investing Amid Low Expected Returns: Making the Most When Markets Offer the Least
by Antti Ilmanen
Published 24 Feb 2022

I touch upon these issues elsewhere in the book, albeit briefly. Box 10.1 Machine Learning The classic forecasting approaches rely on stable linear relationships between predictors and future returns. Many statistical techniques allow incorporation of various complexities and nonlinearities – tailoring contextual models for different markets or sectors, conditional models for different environments, interaction effects between predictors, and more. Newer techniques, called machine learning or artificial intelligence, do not specify these relationships but let a machine learn any complex relationships between predictors and future returns.

…

Newer techniques, called machine learning or artificial intelligence, do not specify these relationships but let a machine learn any complex relationships between predictors and future returns. The main departure of machine learning from its statistical roots is in the massively increased computing power. Machine learning also allows the use of much larger data sets (“big data”), including unstructured data such as text or images. Downsides include heightened overfitting risks and black-box nature. Machine learning is already having a huge impact in many fields. In finance too, it shows great promise but may also be overhyped. Machine learning is an evolutionary, not revolutionary, extension of statistical methods long used in quantitative investing. When it comes to forecasting multi-month or multiyear returns (not high-frequency returns), machine learning still faces limitations of small data.

…

When it comes to forecasting multi-month or multiyear returns (not high-frequency returns), machine learning still faces limitations of small data. Worse, the signal-to-noise ratio is inevitably low in financial markets when competition ensures that predicting returns cannot be too easy. Combined with some innovative alternative data sources, machine learning may help predict the next macroeconomic or firm-specific announcements (“nowcasting”) which would give some edge in short-term trading, but this would do nothing to help us collectively in the low expected return challenge. Machine learning may improve stock selection and market timing strategies, but should be aided by theoretical models to limit overfitting.

Inviting Disaster
by James R. Chiles
Published 7 Jul 2008

I suggest that companies think of their time on the machine frontier as a privilege, bestowed by the rest of us. It’s a valuable location because working the frontier offers more opportunity for profit and growth than old technological territory. In our era, some frontiers with seemingly great potential include genetic engineering, artificial intelligence, and deepwater exploration for oil and gas. But holding a place on the machine frontier is not a constitutional right. Losing the privilege could happen if a laboratory lets something loose onto the citizenry or if some cost-cutting decision causes the destruction of a valued habitat. One mishap might not cause public support to collapse, but two incidents could be more than enough.

…

Cooper never visited the project again, relying on mail and telegrams to keep him in touch with the construction of his design, a design that had been forced to the edges of safety by financial difficulties in the Quebec Bridge Company. When signs of trouble began appearing—steel ribs not lining up, because of the higher-than-estimated weight of the structure—Cooper wasn’t there to see it firsthand, as he had been with the Eads Bridge. When the steel began buckling and the alignment problems grew worse by the day, Cooper heard about it but could do no more than send worried telegrams. His last telegram tried to order the crews to add no more weight to the bridge until after investigation, but he’d sent it via the steel fabricator’s factory, and the message didn’t make it to the construction site in time.

Python for Data Analysis
by Wes McKinney
Published 30 Dec 2011

MovieLens 1M Data Set GroupLens Research (http://www.grouplens.org/node/73) provides a number of collections of movie ratings data collected from users of MovieLens in the late 1990s and early 2000s. The data provide movie ratings, movie metadata (genres and year), and demographic data about the users (age, zip code, gender, and occupation). Such data is often of interest in the development of recommendation systems based on machine learning algorithms. While I will not be exploring machine learning techniques in great detail in this book, I will show you how to slice and dice data sets like these into the exact form you need. The MovieLens 1M data set contains 1 million ratings collected from 6000 users on 4000 movies. It’s spread across 3 tables: ratings, user information, and movie information.

…

Data Munging Topics Many helpful data munging tools for financial applications are spread across the earlier chapters. Here I’ll highlight a number of topics as they relate to this problem domain. Time Series and Cross-Section Alignment One of the most time-consuming issues in working with financial data is the so-called data alignment problem. Two related time series may have indexes that don’t line up perfectly, or two DataFrame objects might have columns or row labels that don’t match. Users of MATLAB, R, and other matrix-programming languages often invest significant effort in wrangling data into perfectly aligned forms. In my experience, having to align data by hand (and worse, having to verify that data is aligned) is a far too rigid and tedious way to work.

…

Preparation Cleaning, munging, combining, normalizing, reshaping, slicing and dicing, and transforming data for analysis. Transformation Applying mathematical and statistical operations to groups of data sets to derive new data sets. For example, aggregating a large table by group variables. Modeling and computation Connecting your data to statistical models, machine learning algorithms, or other computational tools Presentation Creating interactive or static graphical visualizations or textual summaries In this chapter I will show you a few data sets and some things we can do with them. These examples are just intended to pique your interest and thus will only be explained at a high level.

HBase: The Definitive Guide
by Lars George
Published 29 Aug 2011

Ganglia and its graphs are a great tool to go back in time and find what caused a problem. However, they are only helpful when dealing with quantitative data—for example, for performing postmortem analysis of a cluster problem. In the next section, you will see how to complement the graphing with a qualitative support system. Figure 10-5. Graphs that can help align problems with related events * * * [112] Ganglia is a distributed, scalable monitoring system suitable for large cluster systems. See its project website for more details on its history and goals. [113] See the RRDtool project website for details. JMX The Java Management Extensions technology is the standard for Java applications to export their status.

…

Because of this, companies have become focused on delivering more targeted information, such as recommendations or online ads, and their ability to do so directly influences their success as a business. Systems like Hadoop[6] now enable them to gather and process petabytes of data, and the need to collect even more data continues to increase with, for example, the development of new machine learning algorithms. Where previously companies had the liberty to ignore certain data sources because there was no cost-effective way to store all that information, they now are likely to lose out to the competition. There is an increasing need to store and analyze every data point they generate. The results then feed directly back into their e-commerce platforms and may generate even more data.

…

Facebook, for example, is adding more than 15 TB of data into its Hadoop cluster every day[9] and is subsequently processing it all. One source of this data is click-stream logging, saving every step a user performs on its website, or on sites that use the social plug-ins offered by Facebook. This is an ideal case in which batch processing to build machine learning models for predictions and recommendations is appropriate. Facebook also has a real-time component, which is its messaging system, including chat, wall posts, and email. This amounts to 135+ billion messages per month,[10] and storing this data over a certain number of months creates a huge tail that needs to be handled efficiently.