Executive Order (EO) 14028, "Improving the Nation's Cybersecurity," 12 May2021, directs the National Institute of Standards and Technology (NIST) torecommend minimum standards for software testing within 60 days. This documentdescribes eleven recommendations for software verifica … | Continue reading
Social media, e-commerce, streaming video, e-mail, cloud documents, webpages, traffic flows, and network packets fill vast digital lakes, rivers, andoceans that we each navigate daily. This digital hyperspace is an amorphousflow of data supported by continuous streams that stretc … | Continue reading
Energy conservation is a basic physics principle, the breakdown of whichoften implies new physics. This paper presents a method for data-driven "newphysics" discovery. Specifically, given a trajectory governed by unknownforces, our Neural New-Physics Detector (NNPhD) aims to dete … | Continue reading
A decade after its first release, the Go programming language has become amajor programming language in the development landscape. While praised for itsclean syntax and C-like performance, Go also contains a strong statictype-system that prevents arbitrary type casting and arbitr … | Continue reading
Mixed Integer Programming (MIP) solvers rely on an array of sophisticatedheuristics developed with decades of research to solve large-scale MIPinstances encountered in practice. Machine learning offers to automaticallyconstruct better heuristics from data by exploiting shared str … | Continue reading
We present a neural video compression method based on generative adversarialnetworks (GANs) that outperforms previous neural video compression methods andis comparable to HEVC in a user study. We propose a technique to mitigatetemporal error accumulation caused by recursive frame … | Continue reading
This discussion applies quantitative finance methods and economic argumentsto cryptocurrencies in general and bitcoin in particular -- as there are about$10,000$ cryptocurrencies, we focus (unless otherwise specified) on the mostdiscussed crypto of those that claim to hew to the … | Continue reading
Ganymede's atmosphere is produced by charged particle sputtering andsublimation of its icy surface. Previous far-ultraviolet observations of theOI1356-A and OI1304-A oxygen emissions were used to derive sputtered molecularoxygen (O$_2$) as an atmospheric constituent but an expect … | Continue reading
Spectre v1 attacks, which exploit conditional branch misprediction, are oftenidentified with attacks that bypass array bounds checking to leak data from avictim's memory. Generally, however, Spectre v1 attacks can exploit anyconditional branch misprediction that makes the victim … | Continue reading
Transient execution attacks that exploit speculation have raised significantconcerns in computer systems. Typically, branch predictors are leveraged totrigger mis-speculation in transient execution attacks. In this work, wedemonstrate a new class of speculation-based attack that … | Continue reading
CUDA (formerly an abbreviation of Compute Unified Device Architecture) is aparallel computing platform and API model created by Nvidia allowing softwaredevelopers to use a CUDA-enabled graphics processing unit (GPU) for generalpurpose processing. This 90-pages tutorial introduces … | Continue reading
BGP-Multipath (BGP-M) is a multipath routing technique for load balancing.Distinct from other techniques deployed at a router inside an Autonomous System(AS), BGP-M is deployed at a border router that has installed multipleinter-domain border links to a neighbour AS. It uses the … | Continue reading
The arithmetic average of a collection of observed values of a homogeneouscollection of quantities is often taken to be the most representativeobservation. There are several arguments supporting this choice the moment ofinertia being the most familiar. But what does this mean? I … | Continue reading
While modern software development heavily uses versioned packages,programming languages rarely support the concept of versions in theirsemantics, which makes software updates more bulky and unsafe. This paperproposes a programming language that intrinsically supports versions. Th … | Continue reading
The security of the Internet rests on a small number of open-sourcecryptographic libraries: a vulnerability in any one of them threatens tocompromise a significant percentage of web traffic. Despite this potential forsecurity impact, the characteristics and causes of vulnerabilit … | Continue reading
Given the increasing promise of Graph Neural Networks (GNNs) in real-worldapplications, several methods have been developed for explaining theirpredictions. So far, these methods have primarily focused on generatingsubgraphs that are especially relevant for a particular predictio … | Continue reading
This paper introduces the PettingZoo library and the accompanying AgentEnvironment Cycle ("AEC") games model. PettingZoo is a library of diverse setsof multi-agent environments with a universal, elegant Python API. PettingZoowas developed with the goal of accelerating research in … | Continue reading
Specifying the semantics of a programming language formally can have manybenefits. However, it can also require a huge effort. The effort can besignificantly reduced by translating language syntax to so-called fundamentalconstructs (funcons). A translation to funcons is easy to u … | Continue reading
We consider a multiproduct monopoly pricing model. We provide sufficientconditions under which the optimal mechanism can be implemented via upgradepricing -- a menu of product bundles that are nested in the strong set order.Our approach exploits duality methods to identify condit … | Continue reading
We introduce the $L_p$ Traveling Salesman Problem ($L_p$-TSP), given by anorigin, a set of destinations, and underlying distances. The objective is toschedule a destination visit sequence for a traveler of unit speed to minimizethe Minkowski $p$-norm of the resulting vector of vi … | Continue reading
Spatial safety violations are the root cause of many security attacks andunexpected behavior of applications. Existing techniques to enforce spatialsafety work broadly at either object or pointer granularity. Object-basedapproaches tend to incur high CPU overheads, whereas pointe … | Continue reading
Interpretability in machine learning (ML) is crucial for high stakesdecisions and troubleshooting. In this work, we provide fundamental principlesfor interpretable ML, and dispel common misunderstandings that dilute theimportance of this crucial topic. We also identify 10 technic … | Continue reading
The search for extraterrestrial intelligence (SETI) is a scientific endeavorwhich struggles with unique issues -- a strong indeterminacy in what data tolook for and when to do so. This has led to attempts at finding bothfundamental limits of the communication between extraterrest … | Continue reading
The recent emergence of deepfakes, computerized realistic multimedia fakes,brought the detection of manipulated and generated content to the forefront.While many machine learning models for deepfakes detection have been proposed,the human detection capabilities have remained far … | Continue reading
Using a bicycle for commuting is still uncommon in US cities, although itbrings many benefits to both the cyclists and to society as a whole. Cyclinghas the potential to reduce traffic congestion and emissions, increasemobility, and improve public health. To convince people to co … | Continue reading
Fixed-point iterations are at the heart of numerical computing and are oftena computational bottleneck in real-time applications, which typically insteadneed a fast solution of moderate accuracy. Classical acceleration methods forfixed-point problems focus on designing algorithms … | Continue reading
This book develops an effective theory approach to understanding deep neuralnetworks of practical relevance. Beginning from a first-principlescomponent-level picture of networks, we explain how to determine an accuratedescription of the output of trained networks by solving layer … | Continue reading
We present a new dataset of Wikipedia articles each paired with a knowledgegraph, to facilitate the research in conditional text generation, graphgeneration and graph representation learning. Existing graph-text paireddatasets typically contain small graphs and short text (1 or f … | Continue reading
Optimizing resource allocation for analytical workloads is vital for reducingcosts of cloud-data services. At the same time, it is incredibly hard for usersto allocate resources per query in serverless processing systems, and theyfrequently misallocate by orders of magnitude. Unf … | Continue reading
Delivering malware covertly and detection-evadingly is critical to advancedmalware campaigns. In this paper, we present a method that delivers malwarecovertly and detection-evadingly through neural network models. Neural networkmodels are poorly explainable and have a good genera … | Continue reading
The presence of source code that is no longer needed is a handicap to projectmaintainability. The larger and longer-lived the project, the higher thechances of accumulating dead code in its different forms. Manually detecting unused code is time-consuming, tedious, error-prone, … | Continue reading
Blockchain analysis is essential for understanding how cryptocurrencies likeBitcoin are used in practice, and address clustering is a cornerstone ofblockchain analysis. However, current techniques rely on heuristics that havenot been rigorously evaluated or optimized. In this pap … | Continue reading
With the wide application of deep neural networks, it is important to verifya host's possession over a deep neural network model and protect the model. Tomeet this goal, various mechanisms have been designed. By embedding extrainformation into a network and revealing it afterward … | Continue reading
Candidates arrive sequentially for an interview process which results in thembeing ranked relative to their predecessors. Based on the ranks available ateach time, one must develop a decision mechanism that selects or dismisses thecurrent candidate in an effort to maximize the ch … | Continue reading
Serverless computing, or Function-as-a-Service (FaaS), enables a new way ofbuilding and scaling applications by allowing users to deploy fine-grainedfunctions while providing fully-managed resource provisioning and auto-scaling.Custom FaaS container support is gaining traction as … | Continue reading
In the spirit of Trimble's ``Astrophysics in XXXX'' series, I very brieflyand subjectively review developments in SETI in 2020. My primary focus is 74papers and books published or made public in 2020, which I sort into six broadcategories: results from actual searches, new search … | Continue reading
Top-performance sources of photonic entanglement are an indispensableresource for many applications in quantum communication, most notably quantumkey distribution. However, up to now, no source has been shown tosimultaneously exhibit the high pair-creation rate, broad bandwidth, … | Continue reading
The largest store of continually updating knowledge on our planet can beaccessed via internet search. In this work we study giving access to thisinformation to conversational agents. Large language models, even though theystore an impressive amount of knowledge within their weigh … | Continue reading
We model the quantities appearing in Internal Revenue Service (IRS) taxguidance for calculating the health insurance premium tax credit created by thePatient Protection and Affordable Care Act, also called Obamacare. We ask thequestion of whether there is a procedure, computable … | Continue reading
We introduce the GANformer, a novel and efficient type of transformer, andexplore it for the task of visual generative modeling. The network employs abipartite structure that enables long-range interactions across the image,while maintaining computation of linear efficiency, that … | Continue reading
We consider repair tasks: given a critic (e.g., compiler) that assesses thequality of an input, the goal is to train a fixer that converts a bad example(e.g., code with syntax errors) into a good one (e.g., code with no syntaxerrors). Existing works create training data consistin … | Continue reading
A method is proposed, called channel polarization, to construct codesequences that achieve the symmetric capacity $I(W)$ of any given binary-inputdiscrete memoryless channel (B-DMC) $W$. The symmetric capacity is the highestrate achievable subject to using the input letters of th … | Continue reading
The world of empirical machine learning (ML) strongly relies on benchmarks inorder to determine the relative effectiveness of different algorithms andmethods. This paper proposes the notion of "a benchmark lottery" that describesthe overall fragility of the ML benchmarking proces … | Continue reading
Rust, an emerging programming language with explosive growth, provides arobust type system that enables programmers to write memory-safe and data-racefree code. To allow access to a machine's hardware and to support low-levelperformance optimizations, a second language, Unsafe Ru … | Continue reading
Structured pruning is a commonly used technique in deploying deep neuralnetworks (DNNs) onto resource-constrained devices. However, the existingpruning methods are usually heuristic, task-specified, and require an extrafine-tuning procedure. To overcome these limitations, we prop … | Continue reading
We present practical approaches of using deep learning to create and enhancelevel maps and textures for video games -- desktop, mobile, and web. We aim topresent new possibilities for game developers and level artists. The task ofdesigning levels and filling them with details is … | Continue reading
The rise of streaming libraries such as Akka Stream, Reactive Extensions, andLINQ popularized the declarative functional style of data processing. Thestream paradigm offers concise syntax to write down processing pipelines toconsume the vast amounts of real-time data available to … | Continue reading