Adaptive Group Testing on Networks with Community Structure

Theory and simulations show that when the network and infection parametersare conducive to strong community structure, our proposed adaptive,graph-aware algorithm outperforms the baseline binary splitting algorithm, and is even order-optimal in certain parameter regimes . Authors: We derive novelinformation-theoretic lower bounds which highlight the fundamental limits of adaptive group testing in our networked setting.…

Dataset Definition Standard DDS

This document gives a set of recommendations to build and manipulate the datasets used to develop and/or validate machine learning models such as deepneural networks . This is a work in progress as good practicesevolve along with our understanding of machine learning .…

A generalization of the Von Neumann extractor

An iterative randomness extraction algorithm which generalized the VonNeumann’s extraction algorithm is detailed, analyzed and implemented instandard C++ . Given a sequence of independently and identically distributedbiased Bernoulli random variables, to extract randomness from theaforementioned sequence pertains to produce a new sequence of individually andidentically distributed unbiased random variables .…

Towards Optimally Efficient Tree Search with Deep Temporal Difference Learning

This paper investigates the classical integer least-squares problem whichestimates integer signals from linear models . Problem is NP-hard and oftenarises in diverse applications such as signal processing, bioinformatics, communications and machine learning . We propose a generalhyper-accelerated tree search (HATS) algorithm by employing a deep neuralnetwork to estimate the optimal heuristic for the underlying simplified memory-bounded A* algorithm .…

Simplified DOM Trees for Transferable Attribute Extraction from the Web

SimpDOM outperforms the state-of-the-art (SOTA) method by 1.44% on the F1score. We also find that utilizing knowledge from a different vertical(cross-vertical extraction) is surprisingly useful and helps beat the SOTA by afurther 1.37% . The method is transferable and efficient to efficiently retrieve useful context for each node by leveraging the structure of the HTML DOM tree structure .…

The Geometry of the space of Discrete Coalescent Trees

Computational inference of dated evolutionary histories relies upon varioushypotheses about RNA, DNA, and protein sequence mutation rates . Using mutationrates to infer these dated histories is referred to as molecular clockassumption . Coalescent theory is a popular class of evolutionary models thatimplements the molecular clock hypothesis to facilitate computational inferenceof dated phylogenies .…

Distributed Quantum Faithful Simulation and Function Computation Using Algebraic Structured Measurements

In this work, we consider the task of faithfully simulating a distributedquantum measurement and function computation, and demonstrate a new achievableinformation-theoretic rate-region . For this, we develop the technique ofrandomly generating structured POVMs using algebraic codes . We develop a Pruning Traceinequality which is a tighter version of the known operator Markov inequality .…

Merging with unknown reliability

Merging beliefs depends on the relative reliability of their sources . Whenunknown, assuming equal reliability is unwarranted . Alternatively, one source is completely reliable,but which one is unknown . These two cases motivate two existing forms of merging: maxcons-based merging and arbitration .…

Complexity Growth in Integrable and Chaotic Models

We use the SYK family of models with $N$ Majorana fermions to study the complexity of time evolution . Initially, the shortest geodesic followsthe time evolution trajectory, and hence complexity grows linearly in time . Westudy how this linear growth is eventually truncated by the appearance andaccumulation of conjugate points, which signal the presence of shortergeodesics intersecting time evolution trajectories .…

Fine Grained Complexity of Regular Path Queries

A regular path query (RPQ) is a regular expression that returns all nodepairs (u, v) from a graph database that are connected by an arbitrary path labelled with a word from L(q) A main insight isthat we can achieve optimal (or near optimal) algorithms with the PG-approach, but the delay for enumeration is rather high (linear in the database) We explore three successful approaches towards enumeration with sub-linear delay:super-linear preprocessing, approximations of the solution sets, and restricted classes of RPQs .…

FLGUARD Secure and Private Federated Learning

Recently, federated learning (FL) has been subject to both security and privacy attacks posing a dilemmatic challenge on the underlying algorithmicdesigns . FLGUARD is a novel in-depth defense for FL that tackles this challenge . It applies a multilayered defense by using a Model Filtering layer to detect and reject malicious model updates and a Poison Elimination layer to eliminate any effect of a remaining weak manipulation .…

RANK AI assisted End to End Architecture for Detecting Persistent Attacks in Enterprise Networks

Advanced Persistent Threats (APTs) are sophisticated multi-step attacks,planned and executed by skilled adversaries . Intrusion Detection Systems (IDSs) and User and EntityBehavior Analytics (UEBA) are commonly employed to aid a security analyst inthe detection of APTs . In this paper, we provide, up to our knowledge, the first study and implementation of anend-to-end AI-assisted architecture for detecting APTs — RANK .…

Large Scale Extended Granger Causality for Classification of Marijuana Users From Functional MRI

It has been shown in the literature that marijuana use is associated with changes in brain network connectivity . We propose large-scale Extended GrangerCausality (lsXGC) and investigate whether it can capture such changes usingresting-state fMRI . We investigate whether this model can serve as abiomarker for classifying marijuana users from typical controls using 126 adults with a childhood diagnosis of ADHD from the Addiction ConnectomePreprocessed Initiative (ACPI) database .…

Positive first order logic on words

We study FO+ a fragment of first-order logic on finite words . We show that there is a FO-definablelanguage that is monotone in monadic predicates but not definable in FO+. This provides a simple proof that Lyndon’s preservation theorem fails on finitestructures .…

Optimal Action based or User Prediction based Haptic Guidance Can You Do Even Better

Haptic guidance (HG) improves users’ task performancethrough physical interaction between robots and users . We propose implementation methods for each HG type using deeplearning-based approaches . UPHG induces better subjective evaluations, such as naturalness and comfort, than OAHG . CombHG that we proposed furtherdecreases disagreement between the user intention and HG, without reducing the objective and subjective scores.…

Taxonomy Completion via Triplet Matching Network

Triplet Matching Network (TMN) finds the appropriate pairs for a given query concept . TMN consists of one primal scorer and multiple auxiliary scorers . An innovative channel-wisegating mechanism that retains task-specific information in concept representations is introduced to further boost model performance .…

Aerial Ground Interference Mitigation for Cellular Connected UAV

To support large-scale deployment of unmanned aerial vehicles (UAVs) infuture, a new wireless communication paradigm, namely, cellular-connected UAV, has recently received an upsurge of interests in both academia and industry . The highaltitude of UAVs results in more frequent line-of-sight (LoS) channels with both their associated and non-associated BSs in a much wider area, which causes aerial-ground interference .…

Impact of Inter Channel Interference on Shallow Underwater Acoustic OFDM Systems

This paper investigates the impacts of Inter-Channel Interference (ICI)effects on a shallow underwater acoustic (UWA) orthogonal frequency-divisionmultiplexing (OFDM) communication system . The channel capacity of aUWA system is severely suffered by the ICI effect . Numerical results show that the various factors of a UWA-OFDM system as subcarriers, bandwidth, andOFDM symbols affect the channel capacity under the different Dopplerfrequencies .…

One shot Policy Elicitation via Semantic Reward Manipulation

Single-shot Policy Explanation for Augmenting Rewards (SPEAR) is an algorithm that uses semantic explanations derivedfrom combinations of planning predicates to augment agents’ reward functions . SPEAR makes substantial improvements over the current state-of-the-art in terms of runtime and addressable problem size, according to the authors .…

Dynamic Prioritization for Conflict Free Path Planning of Multi Robot Systems

Planning collision-free paths for multi-robot systems (MRS) is a challenging problem because of the safety and efficiency constraints required for real-world solutions . While mostcentrally-planned algorithms use static prioritization, a dynamicprioritization algorithm, PD* is proposed that employs a novel metric, calledfreedom index, to decide the priority order of the robots at each time step .…

Logics of First Order Constraints A Category Independent Approach

We present a general, category independent approach to Logics ofFirst-Order Constraints (LFOC) Traditional First-Order Logic, DescriptionLogic and the sketch framework are discussed as examples . We use the concept ofinstitution [Diaconescu08,GoguenBurstall92] as a guideline to describe LFOC’s . The main result states that any choice of the six parameters, we are going to describe, gives us a corresponding “institution of constraints” at hand .…

Multi Cell Multi Channel URLLC with Probabilistic Per Packet Real Time Guarantee

Ultra-reliable, low-latency communication (URLLC) represents a new focus in 5G-and-beyond networks . It is expected to enable mission-critical sensing and control as well as AR/VR applications . We propose a real-time scheduling algorithm based on local-deadline-partition (LDP) The LDP algorithm is suitable for distributed implementation, and it ensures probabilistic per-packet real time guarantees formulti-cell, multi-channel networks with general deadline constraints .…

Hypothesis Stitcher for End to End Speaker attributed ASR on Long form Multi talker Recordings

An end-to-end (E2E) speaker-attributed automatic speech recognition (SA-ASR) model was proposed recently to jointly perform speaker counting, speechrecognition and speaker identification . However, the E2E model is susceptible to the mismatch between the training and testing conditions . The proposed method significantlyimproves SA-WER especially for long-form multi-talker recordings, especially for multi-talkinger recordings .…

Confluence up to Garbage in Graph Transformation

The transformation of graphs and graph-like structures is ubiquitous incomputer science . When a system is described by graph-transformation rules, it is often desirable that the rules are both terminating and confluent . However, there are application scenarios where rules are not globallyconfluent but confluent on a subclass of graphs that are of interest .…