THE ALIGNMENT PROBLEM: Machine Learning and Human Values by Brian Christian. New York: W. W. Norton, 2020. 344 pages. Hardcover; $28.95. ISBN: 9780393635829. *The global conversation about artificial intelligence (AI) is increasingly polemic--"AI will change the world!" "AI will ruin the world!" Amidst the strife, Brian Christian's work stands out. It is thoughtful, nuanced, and, at times, even poetic. Coming on the heels of his two other bestsellers, The Most Human Human and Algorithms to Live By, this meticulously researched recounting of the last decade of research into AI safety provides a broad perspective of the field and its future. *The "alignment problem" in the title refers to the disconnect between what AI does and what we want it to do. In Christian's words, it is the disconnect between "machine learning and human values." This disconnect has been the subject of intense research in recent years, as both companies and academics continually discover that AIs inherit the mistakes and biases of their creators. *For example, we train AIs that predict recidivism rates of convicted criminals in hopes of crafting more accurate sentences. However, the AIs produce racially biased outcomes. Or, we train AIs which map words into mathematical spaces. These AIs can perform mathematical "computations" on words, such as "king - man + woman = queen" and "Paris - France + Italy = Rome." But they also say that "doctor - man + woman = nurse" and "computer programmer - man + woman = homemaker." These examples of racial and gender bias are some of the numerous ways that human bias appears inside the supposedly impartial tools we have created. *As Norbert Wiener, a famous mathematician in the mid-twentieth century, put it, "We had better be sure the purpose put into the machine is the purpose which we really desire" (p. 312). The discoveries of the last ten years have shocked researchers into realizing that our machines have purposes we never intended. Christian's message is clear: these mistakes must be fixed before those machines become a fixed part of our everyday lives. *The book is divided into three main sections. The first, Prophecy, provides a historical overview of how researchers uncovered the AI biases that are now well known. It traces the origins of how AI models ended up in the public sphere and the history of how people have tried to solve the problems AI creates. Perhaps one of the most interesting anecdotes in this section is about how researchers try to create explainable models to comply with GDPR requirements. *The second section, Agency, explores the alignment problem in the context of reinforcement learning. Reinforcement learning involves teaching computer "agents" (aka AIs) to perform certain tasks using complex reward systems. Time and time again, the reward systems that researchers create have unintended side effects, and Christian recounts numerous humorous examples of this. He explains in simple terms why it is so difficult to correctly motivate the behaviors we wish to see in others (both humans and machines), and what it might take to create machines which are truly curious. This section feels a bit long. Christian dives deeply into the research of a few specific labs and appears to lose his logical thread in the weeds of research. Eventually, he emerges. *The final section, Normativity, provides perspective on current efforts to understand and fix the alignment problem. Its subchapters, "Imitation," "Inference," and "Uncertainty," reference different qualities that human researchers struggle to instill in machines. Imitating correct behaviors while ignoring bad ones is hard, as is getting a machine to perform correctly on data it hasn't seen before. Finally, teaching a model (and humans reading its results) to correctly interpret uncertainty is an active area of research with no concrete solutions. *After spending over three hundred pages recounting the pitfalls of AI and the difficulties of realigning models with human values, Christian ends on a hopeful note. He postulates that the issues discovered in machine-learning models illuminate societal issues that might otherwise be ignored. "Unfair pretrial detection models, for one thing, shine a spotlight on upstream inequities. Biased language models give us, among other things, a way to measure the state of our discourse and offer us a benchmark against which to try to improve and better ourselves ... In seeing a kind of mind at work as it digests and reacts to the world, we will learn something both about the world and also, perhaps, about minds" (p. 328). *As a Christ-follower, I believe the biases found in AI are both terrible and unsurprising. Humans are imperfect creators. While researchers' efforts to fix biases and shortcomings in AI systems are important and worthwhile, they can never exorcise fallen human nature from AI. Christian's conclusions about AI pointing to biases in humans comes close to this idea but avoids taking an overtly theological stance. *This book is well worth reading for those who wish to better understand the limitations of AI and current efforts to fix them. It weaves together history, mathematics, ethics, and philosophy, while remaining accessible to a broad audience through smooth explanations of detailed concepts. You don't need to be an AI expert (or even familiar with AI at all) to appreciate this book's insights. *After you're done reading it, recommend this book to the next person who tells you, with absolute certainty, that AI will either save or ruin the world. Christian's book provides a much-needed dose of sanity and perspective amidst the hype. *Reviewed by Emily Wenger, graduate student in the Department of Computer Science, University of Chicago, Chicago, IL 60637.

A potent anti-vascular endothelial growth factor (VEGF) biologic and a compatible delivery system were co-evaluated for protection against wet age-related macular degeneration (AMD) over a 6month period following a single intravitreal (IVT) injection. The anti-VEGF molecule is dimeric, containing two different anti-VEGF domain antibodies (dAb) attached to a human IgG1 Fc region: a dual dAb. The delivery system is based on microparticles of PolyActive™ hydrogel co-polymer. The molecule was evaluated both in vitro for potency against VEGF and in ocular VEGF-driven efficacy modelsin vivo. The dual dAb is highly potent, showing a lower IC50 than aflibercept in VEGF receptor binding assays (RBAs) and retaining activity upon release from microparticles over 12 months in vitro. Microparticles released functional dual dAb in rabbit and primate eyes over 6 months at sufficient levels to protect Cynomolgus against laser-induced grade IV choroidal neovascularisation (CNV). This demonstrates proof of concept for delivery of an anti-VEGF molecule within a sustained-release system, showing protection in a pre-clinical primate model of wet AMD over 6 months. Polymer breakdown and movement of microparticles in the eye may limit development of particle-based approaches for sustained release after IVT injection.

No abstract

Normal human urinary tract epithelial cells (HUC) were neoplastically transformed in vitro using a step-wise strategy. First, a partially transformed non-virus-producing cell line was obtained after infection of HUC with simian virus 40 (SV40). This cell line (SV-HUC-1) was demonstrated to be clonal in origin, as 100% of cells contained at least five of seven marker chromosomes. Marker chromosomes were formed by balanced translocations resulting in a 'pseudodiploid' cell line. SV-HUC-1 showed altered growth properties in vitro (e.g. anchorage independent growth) but failed to form tumors in athymic nude mice, even after 3 years in culture (80 passages). In the studies reported here, SV-HUC-1 at early passages (P15-P19) were exposed to 3-methylcholanthrene (MCA) in three separate experiments. After a six-week post-treatment period of cell culture, cells were inoculated s.c. into athymic nude mice. In all experiments, MCA-treated SV-HUC-1 formed carcinomas in mice usually with a latent period of 5-8 weeks. These carcinomas showed heterogeneity with respect to histopathologies and growth properties in the mice and karyotypes. All the tumors retained SV-HUC-1 chromosome markers, but each independent transformant was aneuploid and contained unique new marker chromosomes. Chromosomes usually altered in tumor cells included numbers 3, 5, 6, 9, 11 and 13. Mutations in the ras family of cellular proto-oncogenes resulting in altered mobility of the p21 protein product were not detected in six cell lines established from independently derived tumors. It is not yet known whether other cellular proto-oncogenes are activated in these tumorigenic transformants. Neither control SV-HUC-1 (which were not exposed to MCA), nor early passage HUC exposed to MCA formed tumors when inoculated into mice. Thus, the tumorigenic transformation of HUC resulted from the combined actions of SV40 and MCA.

To evaluate the efficacy of systemic and intravitreous administration of VEGF Trap (aflibercept) in a nonhuman primate model of choroidal neovascularization (CNV). Methods: VEGF Trap treatment on laser-induced CNV was evaluated in 48 adult cynomolgus monkeys. In the prevention arms of the study, VEGF Trap was administered by intravenous injection (3 or 10 mg/kg weekly) or intravitreous injection (50, 250, or 500 µg/eye every 2 weeks) beginning before laser injury. In the treatment arm, a single intravitreous injection (500 µg) was given 2 weeks following laser injury. Laser-induced lesions were scored from grade 1 (no hyperfluorescence) to grade 4 (clinically relevant leakage). Representative lesions were evaluated histologically. Results: Grade 4 leakage developed at 32.4% and 45.4% of the laser sites in animals receiving intravitreous or intravenous administration of placebo at 2 weeks following laser injury, respectively. In contrast, the development of grade 4 lesions was completely or nearly completely prevented in all groups receiving intrave