You’ve asked yourself this question again and again. Now you have the chance to see inside the minds of angry and controlling men—and change your life. In Why Does He Do That? you will learn about:

• The early warning signs of abuse
• The nature of abusive thinking
• Myths about abusers
• Ten abusive personality types
• The role of drugs and alcohol
• What you can fix, and what you can’t
• And how to get out of an abusive relationship safely

Recently, information retrieval has seen the emergence of dense retrievers, using neural networks, as an alternative to classical sparse methods based on term-frequency. These models have obtained state-of-the-art results on datasets and tasks where large training sets are available. However, they do not transfer well to new applications with no training data, and are outperformed by unsupervised term-frequency methods such as BM25. In this work, we explore the limits of contrastive learning as a way to train unsupervised dense retrievers and show that it leads to strong performance in various retrieval settings. On the BEIR benchmark our unsupervised model outperforms BM25 on 11 out of 15 datasets for the Recall@100. When used as pre-training before fine-tuning, either on a few thousands in-domain examples or on the large MS~MARCO dataset, our contrastive model leads to improvements on the BEIR benchmark. Finally, we evaluate our approach for multi-lingual retrieval, where training data is even scarcer than for English, and show that our approach leads to strong unsupervised performance. Our model also exhibits strong cross-lingual transfer when fine-tuned on supervised English data only and evaluated on low resources language such as Swahili. We show that our unsupervised models can perform cross-lingual retrieval between different scripts, such as retrieving English documents from Arabic queries, which would not be possible with term matching methods.

Available at: https://research.google/pubs/pub46555/ Creating reliable, production-level machine learning systems brings on a host of concerns not found in small toy examples or even large offline research experiments. Testing and monitoring are key considerations for ensuring the production-readiness of an ML system, and for reducing technical debt of ML systems. But it can be difficult to formulate specific tests, given that the actual prediction behavior of any given model is difficult to specify a priori. In this paper, we present 28 specific tests and monitoring needs, drawn from experience with a wide range of production ML systems to help quantify these issues and present an easy to follow road-map to improve production readiness and pay down ML technical debt.

Available at:
https://ieeexplore.ieee.org/document/8804457

Recent advances in machine learning have stimulated widespread interest within the Information Technology sector on integrating AI capabilities into software and services. This goal has forced organizations to evolve their development processes. We report on a study that we conducted on observing software teams at Microsoft as they develop AI-based applications. We consider a nine-stage workflow process informed by prior experiences developing AI applications (e.g., search and NLP) and data science tools (e.g. application diagnostics and bug reporting). We found that various Microsoft teams have united this workflow into preexisting, well-evolved, Agile-like software engineering processes, providing insights about several essential engineering challenges that organizations may face in creating large-scale AI solutions for the marketplace. We collected some best practices from Microsoft teams to address these challenges. In addition, we have identified three aspects of the AI domain that make it fundamentally different from prior software application domains: 1) discovering, managing, and versioning the data needed for machine learning applications is much more complex and difficult than other types of software engineering, 2) model customization and model reuse require very different skills than are typically found in software teams, and 3) AI components are more difficult to handle as distinct modules than traditional software components - models may be "entangled" in complex ways and experience non-monotonic error behavior. We believe that the lessons learned by Microsoft teams will be valuable to other organizations.

Ten laws of simplicity for business, technology, and design that teach us how to need less but get more.

Finally, we are learning that simplicity equals sanity. We're rebelling against technology that's too complicated, DVD players with too many menus, and software accompanied by 75-megabyte "read me" manuals. The iPod's clean gadgetry has made simplicity hip. But sometimes we find ourselves caught up in the simplicity paradox: we want something that's simple and easy to use, but also does all the complex things we might ever want it to do. In The Laws of Simplicity , John Maeda offers ten laws for balancing simplicity and complexity in business, technology, and design—guidelines for needing less and actually getting more.

Machine learning offers a fantastically powerful toolkit for building useful complexprediction systems quickly. This paper argues it is dangerous to think ofthese quick wins as coming for free. Using the software engineering frameworkof technical debt, we find it is common to incur massive ongoing maintenancecosts in real-world ML systems. We explore several ML-specific risk factors toaccount for in system design. These include boundary erosion, entanglement,hidden feedback loops, undeclared consumers, data dependencies, configurationissues, changes in the external world, and a variety of system-level anti-patterns.

Paper available at NIPS:

https://papers.nips.cc/paper/2015/hash/86df7dcfd896fcaf2674f757a2463eba-Abstract.html

We propose Make-A-Video -- an approach for directly translating the tremendous recent progress in Text-to-Image (T2I) generation to Text-to-Video (T2V). Our intuition is simple: learn what the world looks like and how it is described from paired text-image data, and learn how the world moves from unsupervised video footage. Make-A-Video has three advantages: (1) it accelerates training of the T2V model (it does not need to learn visual and multimodal representations from scratch), (2) it does not require paired text-video data, and (3) the generated videos inherit the vastness (diversity in aesthetic, fantastical depictions, etc.) of today's image generation models. We design a simple yet effective way to build on T2I models with novel and effective spatial-temporal modules. First, we decompose the full temporal U-Net and attention tensors and approximate them in space and time. Second, we design a spatial temporal pipeline to generate high resolution and frame rate videos with a video decoder, interpolation model and two super resolution models that can enable various applications besides T2V. In all aspects, spatial and temporal resolution, faithfulness to text, and quality, Make-A-Video sets the new state-of-the-art in text-to-video generation, as determined by both qualitative and quantitative measures.

The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train. Our model achieves 28.4 BLEU on the WMT 2014 English-to-German translation task, improving over the existing best results, including ensembles by over 2 BLEU. On the WMT 2014 English-to-French translation task, our model establishes a new single-model state-of-the-art BLEU score of 41.8 after training for 3.5 days on eight GPUs, a small fraction of the training costs of the best models from the literature. We show that the Transformer generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.

This document aims to be a self-contained, mathematically precise overview of transformer architectures and algorithms (*not* results). It covers what transformers are, how they are trained, what they are used for, their key architectural components, and a preview of the most prominent models. The reader is assumed to be familiar with basic ML terminology and simpler neural network architectures such as MLPs.

Trained machine learning models are increasingly used to perform high-impact tasks in areas such as law enforcement, medicine, education, and employment. In order to clarify the intended use cases of machine learning models and minimize their usage in contexts for which they are not well suited, we recommend that released models be accompanied by documentation detailing their performance characteristics. In this paper, we propose a framework that we call model cards, to encourage such transparent model reporting. Model cards are short documents accompanying trained machine learning models that provide benchmarked evaluation in a variety of conditions, such as across different cultural, demographic, or phenotypic groups (e.g., race, geographic location, sex, Fitzpatrick skin type) and intersectional groups (e.g., age and race, or sex and Fitzpatrick skin type) that are relevant to the intended application domains. Model cards also disclose the context in which models are intended to be used, details of the performance evaluation procedures, and other relevant information. While we focus primarily on human-centered machine learning models in the application fields of computer vision and natural language processing, this framework can be used to document any trained machine learning model. To solidify the concept, we provide cards for two supervised models: One trained to detect smiling faces in images, and one trained to detect toxic comments in text. We propose model cards as a step towards the responsible democratization of machine learning and related AI technology, increasing transparency into how well AI technology works. We hope this work encourages those releasing trained machine learning models to accompany model releases with similar detailed evaluation numbers and other relevant documentation.

Available:

https://openai.com/blog/whisper/

We study the capabilities of speech processing systems trained simply to predict large amounts of transcripts of audio on the internet. When scaled to 680,000 hours of multilingual and multitask supervision, the resulting models generalize well to standard benchmarks and are often competitive with prior fully supervised results but in a zero- shot transfer setting without the need for any fine- tuning. When compared to humans, the models approach their accuracy and robustness. We are releasing models and inference code to serve as a foundation for further work on robust speech processing.

Hilariously picaresque, epic in scope, alive with the poetry and vigor of the American people, Mark Twain's story about a young boy and his journey down the Mississippi was the first great novel to speak in a truly American voice. Influencing subsequent generations of writers -- from Sherwood Anderson to Twain's fellow Missourian, T.S. Eliot, from Ernest Hemingway and William Faulkner to J.D. Salinger -- "Huckleberry Finn," like the river which flows through its pages, is one of the great sources which nourished and still nourishes the literature of America.

Tolstoy's epic masterpiece captures with unprecedented immediacy the broad sweep of life during the Napoleonic wars and the brutal invasion of Russia. Balls and soires, the burning of Moscow, the intrigues of statesmen and generals, scenes of violent battles, the quiet moments of everyday life--all in a work whose extraordinary imaginative power has never been surpassed. The Maudes' translation of Tolstoy's epic masterpiece has long been considered the best English version, and now for the first time it has been revised to bring it fully into line with modern approaches to the text. French passages are restored, Anglicization of Russian names removed, and outmoded expressions updated. A new introduction by Amy Mandelker considers the novel's literary and historical context, the nature of the work, and Tolstoy's artistic and philosophical aims. New, expanded notes provide historical background and identifications, as well as insight into Russian life and society.

Contents: A Day in the Country; Old Age; Kashtanka; Enemies; On the Way; Vanka; La Cigale; Grief; An Inadvertence; The Black Monk; The Kiss; In Exile; A Work of Art; Dreams; A Woman's Kingdom; The Doctor; A Trifling Occurrence; The Hollow; After the Theater; The Runaway; Vierochka; The Steppe; Rothschild's Fiddle.

Considered by some to be the greatest novel ever written, Anna Karenina is Tolstoy's classic tale of love and adultery set against the backdrop of high society in Moscow and Saint Petersburg. A rich and complex masterpiece, the novel charts the disastrous course of a love affair between Anna, a beautiful married woman, and Count Vronsky, a wealthy army officer. Tolstoy seamlessly weaves together the lives of dozens of characters, and in doing so captures a breathtaking tapestry of late-nineteenth-century Russian society.

Hamlet is Shakespeare’s most popular, and most puzzling, play. It follows the form of a “revenge tragedy,” in which the hero, Hamlet, seeks vengeance against his father’s murderer, his uncle Claudius, now the king of Denmark. Much of its fascination, however, lies in its uncertainties. Among them: What is the Ghost—Hamlet’s father demanding justice, a tempting demon, an angelic messenger? Does Hamlet go mad, or merely pretend to? Once he is sure that Claudius is a murderer, why does he not act? Was his mother, Gertrude, unfaithful to her husband or complicit in his murder?

Recognized as one of the major literary works of the twentieth-century, Marcel Proust's monumental seven-volume novel brings together memories of childhood and Parisian society before and during the First World War.

A sprawling work set in a provincial English town, Middlemarch boasts a large cast of characters whose stories interweave against a backdrop of political upheaval.

Instead of choosing to wed a wealthy landowner and settle for a comfortable life, Dorothea Brooke decides to marry Edward Casaubon—a dull scholar who is her senior by a few decades. Dorothea hopes the union will afford her with the opportunity to share in her husband’s intellectual pursuits, but even her best efforts can’t save the disastrous marriage. Meanwhile, idealistic doctor Tertius Lydgate has progressive ideas about the medical field, and he believes the village of Middlemarch is the place that will embrace his beliefs. But when he weds Rosamond Vincy, the mayor’s beautiful daughter, his ideals come into stark contrast with his new bride’s materialism and vanity, dooming their marriage from the start.

Lastly, Rosamond's brother, Fred, is reluctantly destined for the Church. But his childhood sweetheart Mary Garth refuses to accept him until he settles into a more suitable career and one that interests him. But when circumstances lead to Fred losing a sizeable fortune that sets off a calamitous turn of events, he must reevaluate the choices he has made.

Emma is bored and wants excitement and makes a complete mess of her life. In the process she destroys the life of her very provincial husband who also happens to be a good man. The story is tragic and at times frustrating but well worth reading.

After marrying Charles Bovary, an ordinary country doctor, Emma Rouault hopes to have the adventurous life of passion and intrigue that she reads about in books. However, she is disappointed with her husband and her simple life. It is anything but the romantic excitement she longs for. In her desire to fulfill her dreams she travels down a winding path of adultery and deception.

Large text-to-image models achieved a remarkable leap in the evolution of AI, enabling high-quality and diverse synthesis of images from a given text prompt. However, these models lack the ability to mimic the appearance of subjects in a given reference set and synthesize novel renditions of them in different contexts. In this work, we present a new approach for "personalization" of text-to-image diffusion models (specializing them to users' needs). Given as input just a few images of a subject, we fine-tune a pretrained text-to-image model (Imagen, although our method is not limited to a specific model) such that it learns to bind a unique identifier with that specific subject. Once the subject is embedded in the output domain of the model, the unique identifier can then be used to synthesize fully-novel photorealistic images of the subject contextualized in different scenes. By leveraging the semantic prior embedded in the model with a new autogenous class-specific prior preservation loss, our technique enables synthesizing the subject in diverse scenes, poses, views, and lighting conditions that do not appear in the reference images. We apply our technique to several previously-unassailable tasks, including subject recontextualization, text-guided view synthesis, appearance modification, and artistic rendering (all while preserving the subject's key features).

Deep Learning

Deep learning is making waves. At the time of this writing (March 2016), Google’s AlghaGo program just beat 9-dan professional Go player Lee Sedol at the game of Go, a Chinese board game.

Experts in the field of Artificial Intelligence thought we were 10 years away from achieving a victory against a top professional Go player, but progress seems to have accelerated!

While deep learning is a complex subject, it is not any more difficult to learn than any other machine learning algorithm. I wrote this book to introduce you to the basics of neural networks. You will get along fine with undergraduate-level math and programming skill.

All the materials in this book can be downloaded and installed for free. We will use the Python programming language, along with the numerical computing library Numpy. I will also show you in the later chapters how to build a deep network using Theano and TensorFlow, which are libraries built specifically for deep learning and can accelerate computation by taking advantage of the GPU.

Unlike other machine learning algorithms, deep learning is particularly powerful because it automatically learns features. That means you don’t need to spend your time trying to come up with and test “kernels” or “interaction effects” - something only statisticians love to do. Instead, we will let the neural network learn these things for us. Each layer of the neural network learns a different abstraction than the previous layers. For example, in image classification, the first layer might learn different strokes, and in the next layer put the strokes together to learn shapes, and in the next layer put the shapes together to form facial features, and in the next layer have a high level representation of faces.

On top of all this, deep learning is known for winning its fair share Kaggle contests. These are machine learning contests that are open to anyone in the world who are allowed to use any machine learning technique they want. Deep learning is that powerful.

Do you want a gentle introduction to this “dark art”, with practical code examples that you can try right away and apply to your own data? Then this book is for you.

Who is this book NOT for?

Deep Learning and Neural Networks are usually taught at the upper-year undergraduate level. That should give you some idea of the type of knowledge you need to understand this kind of material.

You absolutely need exposure to calculus to understand deep learning, no matter how simple the instructor makes things. Linear algebra would help. I will assume familiarity with Python (although it is an easy language to pick up). You will need to have some concept of machine learning. If you know about algorithms like logistic regression already, this book is perfect for you. If not, you might want to check out my “prerequisites” book, at: http://amzn.com/B01D7GDRQ2

On the other hand, this book is more like a casual primer than a dry textbook. If you are looking for material on more advanced topics, like LSTMs, convolutional neural networks, or reinforcement learning, I have online courses that teach this material, for example: https://www.udemy.com/deep-learning-c...

New libraries like TensorFlow are being updated constantly. This is not an encyclopedia for these libraries (as such a thing would be impossible to keep up to date). In the one (1!!!) month since the book was first published, no less than THREE new wrapper libraries for TensorFlow have been released to make coding deep networks easier.

We introduce Jukebox, a model that generates music with singing in the raw audio domain. We tackle the long context of raw audio using a multi-scale VQ-VAE to compress it to discrete codes, and modeling those using autoregressive Transformers. We show that the combined model at scale can generate high-fidelity and diverse songs with coherence up to multiple minutes. We can condition on artist and genre to steer the musical and vocal style, and on unaligned lyrics to make the singing more controllable.

WINNER OF THREE EISNER AWARDS As Yorick Brown, the last man on Earth, begins to make his way across the country to California, he and his companions are forced to make an unscheduled stop in Marrisville, Ohio—a small town with a big secret.

Collects Y: The Last Man issues #6-10

The adventures of Yorick Brown continue in Y: THE LAST MAN: ONE SMALL STEP, collecting issues #11-17.

A Russian Soyuz capsule is coming down from the International Space Station carrying three passengers: one woman and two men. Could this be the end of Yorick's tenure as last living male?

As the Last Man on Earth and his companions continue to head West, the story takes a detour into the psychological.

In the care of a fellow Culper Ring member, Yorick Brown is forced to confront his tremendous feelings of survivor guilt that lead him to constantly put his life in danger. Once on the road again, the group runs up against a literal roadblock in Arizona, where the female remains of the Sons of Arizona militia have cut the interstate to keep out any vestiges of the U.S. government.

Yorick Brown, the last man on Earth, finally makes it to San Francisco where his unbalanced sister, Hero, finds him seemingly succumbing to the male-killing plague after losing his still-unused engagement ring to the burqa-clad agents of the Setauket Ring. But is the ring really the key to his survival? And what does it have to do with the mysterious Amulet of Helene, which the Setauket leader is determined to take from Agent 355 by any means necessary. Collects issues #24-31.

More than two years after the unexplained plague killed every other male mammal on the planet, the two exceptions - aspiring escape artist Yorick Brown and his pet monkey Ampersand - have traveled across the remains of the U.S. with biochemist Dr. Allison Mann and goverment operative Agent 355 to Dr. Mann's lab in San Francisco, where she managed to finally isolate the source of their immunity.

Unfortunately, that source is within Ampersand's body, which was last seen disappearing under the arm of a mysterious Japanese mercenary. Now, accompanied by 355 and Dr. Mann, the last man on Earth has embarked on a new and even more hazardous journey - following Ampersand's trail across the Pacific, where danger threatens from below as well as above the waves.

Collects issues 32-36

The saga of Yorick Brown, the last man on Earth, continues in PAPER DOLLS. In addition to catching up on the adventures of Yorick's monkey Ampersand (whose body holds the key to stopping the male-killing plague) and telling the origin of Agent 355, PAPER DOLLS chronicles Yorick and 355's search for Yorick's fiancee Beth in Australia -- a search that yields a large dose of unwanted publicity for the Last Man, and deadly consequences for those he cares for!

Collects Y: The Last Man issues #37-#42

WINNER OF THREE EISNER AWARDS. KIMONO DRAGONS brings the hunt for Ampersand-the monkey who could unlock the mystery of the male-killing plague- to its explosive climax, as the last man on Earth and his companions finally reach Japan and discover the truth behind Ampersand's abduction. Collects issues #43-48 of Brian K. Vaughan and Pia Guerra's award-winning Vertigo series.

The saga of Yorick Brown, an unemployed and unmotivated slacker who discovers he is the only male left in the world after a plague of unknown origin instantly kills every mammal with a Y chromosome. Accompanied by his mischievous monkey, Ampersand, and the mysterious Agent 355, Yorick embarks on a transcontinental journey to find his long-lost girlfriend and discover why he is the last man on earth.

This volume of the critically acclaimed series features Yorick and Agent 355 preparing for their ultimate quest to reunite the last man with his lost love, while the person, people or thing behind the disaster that wiped out half of humanity is revealed!

WINNER OF THREE EISNER AWARDS

Featured in THE NEW YORK TIMES and on NPR, Y: THE LAST MAN is the gripping saga of Yorick Brown, an unemployed and unmotivated slacker who discovers he is the only male left in the world after a plague of unknown origin instantly kills every mammal with a Y chromosome. Accompanied by his mischievous monkey, Ampersand, and the mysterious Agent 355, Yorick embarks on a transcontinental journey to find his long-lost girlfriend and discover why he is the last man on earth.

Yorick Brown's long journey through an Earth populated only by women comes to a dramatic, unexpected conclusion in this final volume. Collects issues #55-60 of Brian K. Vaughan and Pia Guerra's award-winning Vertigo series.

"Y" is none other than unemployed escape artist Yorick Brown (his father was a Shakespeare buff), and he's seemingly the only male human left alive after a mysterious plague kills all Y-chromosome carriers on earth.

But why are he and his faithful companion, the often testy male monkey Ampersand, still alive? He sets out to find the answer (and his girlfriend), while running from angry female Republicans (now running the government), Amazon wannabes that include his own sister (seemingly brainwashed), and other threats.

V must come to a decision about Evey, but it's one his young compatriot might not like. Plus, V breaks into a television station with an important message.

Evey Hammond has been taken prisoner! Thrown into a cell and questioned and tortured relentlessly, Evey must search for a way to keep herself alive and her spirit itself from shattering.

Evey feels betrayed by her involvement in Bishop Lilliman's murder and gives V an ultimatum. Meanwhile, the Fingermen come closer to unlocking the mystery of V's identity!

Evey is given a choice: confess and work with her captors or face death. It all leads to the truth of her abduction and the beginning of the final chapters of V's saga!

V continues his crusade against the tyrannical government of Great Britain, bombing another prominent building and attacking a Bishop. As Evey grows more comfortable with the vigilante, she falls deeper into his plot.

The final issue of writer Alan Moore and artist David Lloyd's acclaimed series! One may be able to kill a person, but an idea is bulletproof...

In a post-nuclear world, Great Britain has emerged as a Fascist state. When young Evey Hammond is saved by a mysterious masked stranger, she is pulled into his quest to fight back and overthrow the government.

The clone assassin has been played long enough–
now it’s more than a game.

Bred to kill, Agent 47 is The Agency’s most valuable assassin. So when a competing murder-for-hire organization decides to destroy The Agency, the first person they target for elimination is Agent 47. Tasking someone to off the best hitman in the business is one thing; getting the job done is another. When the attempt falls short, Agent 47 is ordered to track down and kill the culprit who is feeding vital information about The Agency to its enemies.

Agent 47 must follow a bloody trail halfway around the world, fight his way through the streets of Fez, Morocco, and battle slavers deep inside Chad. Then he will discover a shattering truth: If he fails at his mission, the price he’ll pay will be far greater than his own life....

47 and his handler, Diana Burnwood, are a team responsible for some of the most ruthless, efficient and untraceable assassinations across the globe. But before they were operatives of the shadowy International Contracts Association, 47 and Burnwood were on their own separate paths in life - streets lined with blood, bullets and revenge. As 14-year-old Diana begins a mission to track down the people who killed her parents, 47 is struggling to free himself and his friend, 6, from the mysterious Institute which has bred them as killers since they were kids. With a story that stretches from the heartland of America to the Berlin Wall, Lone Wolves will trace the birth of the legendary Hitman and the secret past of the woman behind him who pulls the string and knows just how hard taking a life truly is.

In 1941, Astounding Science Fiction magazine published a short story by a little-known writer named Isaac Asimov. The story was called "Nightfall", and many years later it has long been recognized as a classic, its author a legend. Now, the Gran Master of Science Fiction teams with Robert Silverberg, one of the field's top award-winning authors, to explore and expand an apocalyptic tale that is more spellbinding today than ever before -- Nightfall: The Novel.

Imagine living on a planet with six suns that never experiences Darkness. Imagine never having seen the Stars. Then, one by one the suns start to set, gradually leading into Darkness for the first time ever. Kalgash is a world on the edge of chaos, torn between the madness of religious fanaticism and the unyielding rationalism of scientists. Lurking beneath it all is a collective, instinctual fear of the Darkness. For Kalgash knows only the perpetual light of day; to its inhabitants, a gathering twilight portends unspeakable horror. And only a handful of people on the planet are prepared to face the truth, their six suns are setting all at once for the first time in over two thousand years, signaling the end of civilization as it explodes in the awesome splendor of Nightfall.

Encompassing the psychology of disaster, the tenacity of the human spirit, and, ultimately, the regenerative power of hope, Nightfall is a tale rich in character and suspense that only the unique collaboration of Isaac Asimov and Robert Silverberg could create.