Recent work on Neural Radiance Fields (NeRF) showed how neural networks canbe used to encode complex 3D environments that can be renderedphotorealistically from novel viewpoints. Rendering these... | Continue reading
Active crystals are highly ordered structures that emerge from thenonequilibrium self-organization of motile objects, and have been widelystudied in synthetic and bacterial active matter.... | Continue reading
When experiencing an information need, users want to engage with an expert,but often turn to an information retrieval system, such as a search engine,instead. Classical information retrieval... | Continue reading
Provided significant future progress in artificial intelligence andcomputing, it may ultimately be possible to create multiple Artificial GeneralIntelligences (AGIs), and possibly entire... | Continue reading
The strong performance of vision transformers on image classification andother vision tasks is often attributed to the design of their multi-headattention layers. However, the extent to which... | Continue reading
In this paper we present the first deterministic polynomial time algorithmfor determining the existence of a Hamiltonian cycle and finding a Hamiltoniancycle in general graphs. Our algorithm can... | Continue reading
Neural Cellular Automata (NCA) have shown a remarkable ability to learn therequired rules to "grow" images, classify morphologies, segment images, as wellas to do general computation such as... | Continue reading
Transformers, composed of multiple self-attention layers, hold strongpromises toward a generic learning primitive applicable to different datamodalities, including the recent breakthroughs in... | Continue reading
Modern machine learning models for computer vision exceed humans in accuracyon specific visual recognition tasks, notably on datasets like ImageNet.However, high accuracy can be achieved in many... | Continue reading
Predicting and planning interactive behaviors in complex traffic situationspresents a challenging task. Especially in scenarios involving multiple trafficparticipants that interact densely,... | Continue reading
We show that Transformer encoder architectures can be massively sped up, withlimited accuracy costs, by replacing the self-attention sublayers with simplelinear transformations that "mix" input... | Continue reading
Rust claims to advance industrial programming by bridging the gap betweenlow-level systems programming and high-level application programming. At theheart of the argument that this enables... | Continue reading
Kane and Mertz's 2012 AMS Notices article "Debunking Myths about Gender andMathematics Performance" claims to have debunked the greater male variabilityhypothesis with respect to mathematics... | Continue reading
Mistakes in binary conditions are a source of error in many software systems.They happen when developers use, e.g., < or > instead of <= or >=. Theseboundary mistakes are hard to find and impose... | Continue reading
This article proves the Riemann hypothesis, which states that all non-trivialzeros of the zeta function have a real part equal to 1/2. We inspect in detailthe integral form of the (symmetrized)... | Continue reading
We present an approach to enhancing the realism of synthetic images. Theimages are enhanced by a convolutional network that leverages intermediaterepresentations produced by conventional... | Continue reading
Casting machine learning as a type of search, we demonstrate that theproportion of problems that are favorable for a fixed algorithm is strictlybounded, such that no single algorithm can perform... | Continue reading
The traditional Domain Name System (DNS) lacks fundamental features ofsecurity and privacy in its design. As concerns of privacy increased on theInternet, security and privacy enhancements of... | Continue reading
We describe the new field of mathematical analysis of deep learning. Thisfield emerged around a list of research questions that were not answered withinthe classical framework of learning... | Continue reading
Many practices have been presented in music generation recently. Whilestylistic music generation using deep learning techniques has became the mainstream, these models still struggle to generate... | Continue reading
Bitcoin is a digital currency and payment system based on classicalcryptographic technologies which works without a central administrator such asin traditional currencies. It has long been... | Continue reading
We show that diffusion models can achieve image sample quality superior tothe current state-of-the-art generative models. We achieve this onunconditional image synthesis by finding a better... | Continue reading
Transformer architectures show spectacular performance on NLP tasks and haverecently also been used for tasks such as image completion or imageclassification. Here we propose to use a sequential... | Continue reading
Does consciousness collapse the quantum wave function? This idea was takenseriously by John von Neumann and Eugene Wigner but is now widely dismissed. Wedevelop the idea by combining a... | Continue reading
Semantic understanding of programs is a fundamental problem for programminglanguage processing (PLP). Recent works that learn representations of codebased on pre-training techniques in NLP have... | Continue reading
Controversial understandings of the coronavirus pandemic have turned datavisualizations into a battleground. Defying public health officials,coronavirus skeptics on US social media spent much of... | Continue reading
Isolation is a long-standing challenge of software security. Traditionalprivilege rings and virtual memory are more and more augmented with conceptssuch as capabilities, protection keys, and... | Continue reading
The universal Turing machine is generally considered to be the simplest, mostabstract model of a computer. This paper reports on the discovery of anaccidental arbitrary code execution... | Continue reading
With the advance of deep learning technology, automatic video generation fromaudio or text has become an emerging and promising research topic. In thispaper, we present a novel approach to... | Continue reading
Conceptual abstraction and analogy-making are key abilities underlyinghumans' abilities to learn, reason, and robustly adapt their knowledge to newdomains. Despite of a long history of research... | Continue reading
The structure of the icy shells of ocean worlds is important forunderstanding the stability of their underlying oceans as it controls the rateat which heat can be transported outward and... | Continue reading
With the rising number of IoT devices, the security of such devices becomesincreasingly important. Remote attestation (RA) is a distinct security servicethat allows a remote verifer to reason... | Continue reading
Computer-Aided Design (CAD) applications are used in manufacturing to modeleverything from coffee mugs to sports cars. These programs are complex andrequire years of training and experience to... | Continue reading
Visual content often contains recurring elements. Text is made up of glyphsfrom the same font, animations, such as cartoons or video games, are composedof sprites moving around the screen, and... | Continue reading
We catalog and describe Google's key software engineering practices. | Continue reading
Unexplained periodic fluctuations in the decay rates of Si-32 and Ra-226 havebeen reported by groups at Brookhaven National Laboratory (Si-32), and at thePhysikalisch-Technische-Bundesandstalt... | Continue reading
We present a new universal source code for unlabeled binary and ordinal treesthat achieves asymptotically optimal compression for all tree sources coveredby existing universal codes. At the same... | Continue reading
Convolutional Neural Networks (CNNs) are the go-to model for computer vision.Recently, attention-based networks, such as the Vision Transformer, have alsobecome popular. In this paper we show... | Continue reading
Rapid adoption of the serverless (or Function-as-a-Service, FaaS) paradigm,pioneered by Amazon with AWS Lambda and followed by numerous commercialofferings and open source projects, introduces... | Continue reading
Large-scale pre-trained models like BERT, have obtained a great success invarious Natural Language Processing (NLP) tasks, while it is still a challengeto adapt them to the math-related tasks.... | Continue reading
Why does the Sun shine? Today we know the answer to the question and we alsoknow that earlier answers were quite wrong. The problem of the source of solarenergy became an important part of... | Continue reading
Generating videos from text is a challenging task due to its highcomputational requirements for training and infinite possible answers forevaluation. Existing works typically experiment on... | Continue reading
Decentralized Finance (DeFi), a blockchain powered peer-to-peer financialsystem, is mushrooming. One year ago the total value locked in DeFi systems wasapproximately 700m USD, now, as of April... | Continue reading
Large-scale Pretrained Language Models (PLMs) have become the new paradigmfor Natural Language Processing (NLP). PLMs with hundreds of billionsparameters such as GPT-3 have demonstrated strong... | Continue reading
Over the past couple of years, the growing debate around automated facialrecognition has reached a boiling point. As developers have continued toswiftly expand the scope of these kinds of... | Continue reading
Emacs Lisp (Elisp) is the Lisp dialect used by the Emacs text editor family.GNU Emacs can currently execute Elisp code either interpreted orbyte-interpreted after it has been compiled to... | Continue reading
We demonstrate how by using a reinforcement learning algorithm, the deepcross-entropy method, one can find explicit constructions and counterexamplesto several open conjectures in extremal... | Continue reading