A Gaussian fixed point random walk

In this note, we design a discrete random walk on the real line which takessteps $0, \pm 1$ (and one with steps in $\{\pm 1, 2\}$) We recover linear time algorithms for logarithmicbounds for Koml\'{o}s conjecture in an oblivious online setting .…

Student and Faculty Adviser Insights in an Agile Methodology Integrated Filipino Company Sponsored I T Capstone Program

Theseshowed issues with time management, communication, and competency . Groupsthat excelled exhibited better team coordination and a complete grasp of the Agile methodology . For future implementations, clearer task definition and reducing skill gaps are necessary for better execution . For the future implementation of Agile Methodology, clearer tasks definition and reduced skill gaps necessary for future implementation, according to the authors .…

Burling graphs revisited Part 1 New characterizations

Burling sequence is a sequence of triangle-free graphs of increasingchromatic number . Each of them is isomorphic to the intersection graph of a set of axis-parallel boxes in $R^3$ These graphs were also proved to have othergeometrical representations: intersection graphs of line segments in the plane, and intersection graphs .…

Identity Inference on Blockchain using Graph Neural Network

Identity inference, which aims to make apreliminary inference about account identity, plays a significant role in security . As a common tool, graph mining technique can effectivelyrepresent the interactive information between accounts and be used for identityinference . But existing methods cannot balance scalability and end-to-endarchitecture, resulting high computational consumption and weak featurerepresentation .…

Computing the Union Join and Subset Graph of Acyclic Hypergraphs in Subquadratic Time

We investigate the two problems of computing the union join graph and the subset graph for acyclic hypergraphs . We show that, if the Strong Exponential Time Hypothesis is true, both problems cannot be solved in $\mathcal{O} \bigl( N^2 / \log N + |G| \bigr)$ time for $\alpha$-acyclichypergraphs and any constant $varepsilon 0$ We also present algorithms that solve both problems in $mathcal-O} time for $\gamma$ and $beta$ for interimary hypergraphs, where $|G|$ is the size of the computed graph .…

Engineering Predecessor Data Structures for Dynamic Integer Sets

We present highly optimized data structures for the dynamic predecessorproblem . The task is to maintain a set $S$ of $w$-bit numbers underinsertions, deletions, and predecessor queries . The problem of finding predecessors can beviewed as a generalized form of the membership problem, or as a simple version of the nearest neighbour problem .…

Coalescent Computing

As computational infrastructure extends to the edge, it will increasinglyoffer the same fine-grained resource provisioning mechanisms used in large-scale cloud datacenters . Wireless networking technology will allow service providers to blur the distinction between local and remote resources for commodity computing .…

Hypothesis Formalization Empirical Findings Software Limitations and Design Implications

Data analysis requires translating higher level questions and hypotheses intocomputable statistical models . We find thatanalysts fixated on implementation and shaped their analysis to fit familiar approaches, even if sub-optimal . In an analysis of software tools, we find that tools provide inconsistent, low-level abstractions that may limit the models analysts use to formalize hypotheses .…

Efficient conformer based speech recognition with linear attention

Conformer-based end-to-end automatic speech recognition, whichoutperforms recurrent neural network based ones, has received much attention . The proposed model, named linear attention basedconformer (LAC) can be trained and inferenced jointly with the connectionisttemporal classification objective . Results show that the proposed LAC achieves better performance than 7 recently proposed speech recognition models, and iscompetitive with the state-of-the-art conformer .…

FastS2S VC Streaming Non Autoregressive Sequence to Sequence Voice Conversion

This paper proposes a non-autoregressive extension of our previously proposed sequence-to-sequence (S2S) model-based voice conversion (VC) methods . The student model consists of encoder, decoder, and attentionpredictor . The attention predictor learns to predict attention distributionssolely from source speech along with a target class index with the guidance of those predicted by the teacher model from both source and target speech .…

Hypothesis Formalization Empirical Findings Software Limitations and Design Implications

Data analysis requires translating higher level questions and hypotheses intocomputable statistical models . We find thatanalysts fixated on implementation and shaped their analysis to fit familiar approaches, even if sub-optimal . In an analysis of software tools, we find that tools provide inconsistent, low-level abstractions that may limit the models analysts use to formalize hypotheses .…

Finite Volume Neural Network Modeling Subsurface Contaminant Transport

Finite Volume Neural Network (FINN) is a new approach to data-driven modeling of spatiotemporal physical processes . FINN enables better handling of fluxes between controlvolumes and therefore proper treatment of different types of numerical boundary conditions . The FINN method adopts the numerical structure of the well-known Finite volume Method for handling partialdifferential equations, so that each quantity of interest follows its own conservation law, while it concurrently accommodates learnableparameters.…

Bayesian Optimisation for a Biologically Inspired Population Neural Network

We have used Bayesian Optimisation (BO) to find hyper-parameters in an existing biologically plausible population neural network . The 8-dimensional optimal parameters should be such that the network dynamicssimulate the resting state alpha rhythm (8 – 13 Hz rhythms in brain signals) The best combination of these parameters leads to theneural network’s output power spectral peak being constraint within the alphaband.…

A New Coreset Framework for Clustering

Given a metric space, the $(k,z)$-clustering problem consists of finding $k$centers such that the sum of the of distances raised to the power $z$ of everypoint to its closest center is minimized . This encapsulates the famous$k$-median ($z=1$) and $k$.-means ($z =2$) clustering problems .…

Fairness in Rankings and Recommendations An Overview

We increasingly depend on a variety of data-driven algorithmic systems to assist us in many aspects of life . This has given rise to important concerns regarding the fairness of such systems . In this work, we aim at presenting a toolkit of definitions,models and methods used for ensuring fairness in rankings and recommendations .…

Few shot Image Generation via Cross domain Correspondence

Training generative models on a target domain containing limited examples can easily result in overfitting . In this work, weseek to utilize a large source domain for pretraining and transfer the diversity information from source to target . We propose to preserve therelative similarities and differences between instances in the source via anovel cross-domain distance consistency loss .…

VariTex Variational Neural Face Textures

VariTex is the first method that learns a latent feature space ofneural face textures . We combinethis generative model with a parametric face model and gain explicit controlover head pose and facial expressions . To generate images of complete humanheads, we propose an additive decoder that generates plausible additional details such as hair .…

ShapeMOD Macro Operation Discovery for 3D Shape Programs

ShapeMOD is an algorithm that automatically discovers macros that are useful across large datasets of shape programs . ShapeMOD operates on shape programs expressed in animperative, statement-based language . It is designed to discover macros that make programs more compact by minimizing the number of function calls and freeparameters required to represent an input shape collection .…

Going dark Analysing the impact of end to end encryption on the outcome of Dutch criminal court cases

Main argument put forward by law enforcement is that end-to-end encryption (E2EE) hampers authorities prosecuting criminals who rely on encrypted communication . This statement, however, is not supported by evidence, and therefore not suitable as the sole basis ofpolicymaking . In contrast to what the US attorney general wants us tobelieve, at least the prosecution of cases does not seem hampered by E2EE .…