Language Acquisition Environment for Human Level Artificial Intelligence

Despite recent advances in many application-specific domains, we do not know how to build a human-level artificial intelligence (HLAI) We conjecture thatlearning from others’ experience with the language is the essentialcharacteristic that differentiates human intelligence from the rest . In this environment, there are no explicit definitions of tasks or rewards given when accomplishing those tasks .…

Transient Stability Assessment for Current Constrained and Unconstrained Fault Ride Through in Virtual Oscillator Controlled Converters

Unified virtual oscillator controller (uVOC) inherits the rigorous analytical foundation offered by oscillator based grid-forming (GFM) controllers andenables fast over-current limiting and fault ride-through (FRT) The results demonstrate that uVOCis not constrained by a critical clearing angle unlike droop and virtualsynchronous machine (VSM) type second order controllers .…

Multi Channel Automatic Speech Recognition Using Deep Complex Unet

Front-end module in multi-channel automatic speech recognition (ASR)systems mainly use microphone array techniques to produce enhanced signals innoisy conditions with reverberation and echos . Recently, neural network (NN)based front-end has shown promising improvement over the conventional signalprocessing methods . In this paper, we propose to adopt the architecture of deepcomplex Unet (DCUnet) – a powerful complex-valued Unet-structured speechenhancement model – as the front end of the multi-chamber acoustic model .…

Introduction to Core sets an Updated Survey

In optimization or machine learning problems we are given a set of items,usually points in some metric space, and the goal is to minimize or maximize an objective function over some space of candidate solutions . Traditional algorithms cannot handle modern systems that require parallel real-time computations of infinite distributed streams from sensors such as GPS, audio or video that arrive to a cloud, or networks of weaker devices such as smartphones or robots .…

Visual Diver Face Recognition for Underwater Human Robot Interaction

This paper presents a deep-learned facial recognition method for underwater robots to identify scuba divers . The proposed method is able torecognize divers underwater with faces heavily obscured by scuba masks and breathing apparatus . With the ability tocorrectly recognize divers, autonomous underwater vehicles (AUV) will be able to engage in collaborative tasks with the correct person in human-robot teams and ensure that instructions are accepted from only those authorized to command the robots .…

An Efficient and Scalable Deep Learning Approach for Road Damage Detection

Pavement condition evaluation is essential to time preventative orrehabilitative actions and control distress propagation . This paper introduces adeep learning-based surveying scheme to analyze the image-based distress datain real-time . A database consisting of a diverse population of crack distresstypes such as longitudinal, transverse, and alligator cracks, photographedusing mobile-device is used.…

Strong Data Augmentation Sanitizes Poisoning and Backdoor Attacks Without an Accuracy Tradeoff

Data poisoning and backdoor attacks manipulate victim models by maliciously modifying training data . A recent survey of industry professionals revealed heightened fear in the private sector regarding data poisoning . We find that strong data augmentations, such as mixup and CutMix, can diminish the threat of poisoning and backdoor attacks without trading off performance .…

Privileged Knowledge Distillation for Online Action Detection

Online Action Detection (OAD) in videos is proposed as a per-frame labeling task to address the real-time prediction tasks that can only obtain theprevious and current video frames . We propose Privileged Knowledge Distillation (PKD) which schedules a curriculum learning procedure and inserts auxiliarynodes to the student model, both for shrinking the information gap and improving learning performance .…

Explainable AI for System Failures Generating Explanations that ImproveHuman Assistance in Fault Recovery

The failure of intelligent systems, such as robots, can be inevitable, requiring recoveryassistance from users . In this work, we develop automated, natural languageexplanations for failures encountered during an AI agents’ plan execution . These explanations are developed with a focus of helping non-expert usersunderstand different point of failures to better provide recovery assistance .…

Game Plan What AI can do for Football and What Football can do for AI

The rapid progress in artificial intelligence (AI) and machine learning has opened unprecedented analytics possibilities in various team and individuals sports, including baseball, basketball, and tennis . More recently, AItechniques have been applied to football, due to a huge increase in datacollection by professional teams, increased computational power, and advances in machine learning .…

Small Gain Theorem for Safety Verification of Interconnected Systems

A small-gain theorem in the formulation of barrier function is developed in this work for safety verification of interconnected systems . This result is helpful to verify input-to-state safety (ISSf) of the overall system from the ISSf-barrier function . Also, it can be used to obtain a safety set in a higher dimensional space from the safety sets in two lower dimensional spaces .…

Out of Task Training for Dialog State Tracking Models

Dialog state tracking (DST) suffers from severe data sparsity . In this work, wesuccessfully utilize non-dialog data from unrelated NLP tasks to train dialog state trackers . This opens the door to the abundance of unrelated . NLP . tasks benefit from transfer learning andmulti-task learning, in dialog these methods are limited by the amount of available data and by the specificity of dialog applications .…

TJU DHD A Diverse High Resolution Dataset for Object Detection

Large-scale, rich-diversity, and high-resolution datasets play an important role in developing better object detection methods . The TJU-DHD contains 115,354 high-resolved images (52%images have a resolution of 1624$\times$1200 pixels) and 48% images have aresolution of at least 2,560$1,440 pixels . The dataset contains a rich diversity in season variance, illumination variance, and weathervariance .…

Controllable Emotion Transfer For End to End Speech Synthesis

Emotion embedding space learned from references is a straightforward approach for emotion transfer in encoder-decoder structured emotional text to speech systems . However, the transferred emotion in the synthetic speech is notaccurate and expressive enough with emotion category confusions . We propose a novel approach based on Tacotron .…

Combining Gesture and Voice Control for Mid Air Manipulation of CAD Models in VR Environments

Modeling 3D objects in domains like Computer Aided Design (CAD) is time-consuming and comes with a steep learning curve needed to master the process as well as tool complexities . We designed and implemented a prototype system that leverages the strengths of Virtual Reality (VR) hand gesture recognition in combination with the expressiveness of a voice-based interface for the task of 3D modeling .…

Continuous Emotion Recognition with Spatiotemporal Convolutional Neural Networks

Facial expressions are one of the most powerful ways fordepicting specific patterns in human behavior and describing human emotional state . Even for humans, identifying facial expressions is difficult . Automatic video-based systems for facial expression recognition(FER) have often suffered from variations in expressions among individuals, and from a lack of diverse and cross-culture training datasets .…

heSRPT Parallel Scheduling to Minimize Mean Slowdown

Modern data centers serve workloads which are capable of exploiting parallelism . When a job parallelizes across multiple servers it will completemore quickly, but jobs receive diminishing returns from being allocated additional servers . Because allocating multiple servers to a single job isinefficient, it is unclear how best to allocate a fixed number of servers between many parallelizable jobs .…

On the directed tile assembly systems at temperature 1

We show that a model called directed self-assembly at temperature 1 isunable to do complex computations like the ones of a Turing machine . This model can be seen as a generalization of finite automata to 2D languages . We harmonize the notations between these two articles in orderto clearly solve the directed temperature 1 conjecture .…

Attentional Separation and Aggregation Network for Self supervised Depth Pose Learning in Dynamic Scenes

Self-supervision fromepipolar projection can improve robustness and accuracy of the 3Dperception and localization of vision-based robots . The rigidprojection computed by ego-motion cannot represent all scene points, such aspoints on moving objects, leading to false guidance in these regions . We propose an Attentional Separation-and-AggregationNetwork (ASANet) which can learn to distinguish and extract the scene’s static and dynamic characteristics via the attention mechanism .…

Online Exemplar Fine Tuning for Image to Image Translation

Existing techniques to solve exemplar-based image-to-image translation with convolutional neural networks (CNNs) generally require a training phase to optimize the network parameters on domain-specific and task-specific benchmarks . We propose a novel framework, for the first time, to solve this through an online optimization given an input imagepair, called online exemplar fine-tuning .…

Proposing method to Increase the detection accuracy of stomach cancer based on colour and lint features of tongue using CNN and SVM

The DenseNet network has the highest accuracy compared to other deep architectures . The accuracy of this network for gastric cancer detection reaches 91% which shows the superiority of method in comparison to the state-of-the-art methods . The results show that the proposed method iscorrectly able to identify the area of the tongue as well as the patient’sperson from the non-patient .…

Barycode based GJK Algorithm

In this paper, we present a more efficient GJK algorithm to solve the collision detection and distance query problems in 2D . We propose a new barycode-based sub-distance algorithm that provides a simple and unified condition to determine the minimumsimplex .…

Block Rigidity Strong Multiplayer Parallel Repetition implies Super Linear Lower Bounds for Turing Machines

We prove that a sufficiently strong parallel repetition theorem for a specialcase of multiplayer (multiprover) games implies super-linear lower bounds formulti-tape Turing machines with advice . This is the first connection between parallel repetition and lower bounds for timecomplexity and the first major potential implication of a parallel repetitiontheorem with more than two players .…

On Influencing the Influential Disparity Seeding

Social network platforms have become a crucial medium to disseminate the latest political, commercial, and social information . Users with highvisibility are often selected as seeds to spread information and affect their adoption in target groups . We propose a novel seeding framework, namely Disparity seeding, which aims to maximize information spread while reaching a target user group,e.g.,…

Contextual Fusion For Adversarial Robustness

Mammalian brains handle complex reasoning tasks in a gestalt manner by integrating information from regions of the brain that are specialised to individual sensory modalities . In contrast, deep neural networks are usually designed to process one particular information stream and susceptible to various types of adversarial perturbations .…