Age estimation from facial images is typically cast as a label distributionlearning or regression problem, since aging is a gradual progress. Its mainchallenge is the facial feature space w.r.t.... | Continue reading
Autonomous vehicles and robots require a full scene understanding of theenvironment to interact with it. Such a perception typically incorporatespixel-wise knowledge of the depths and semantic... | Continue reading
We present a new semi-parametric approach to synthesize novel views of anobject from a single monocular image. First, we exploit man-made objectsymmetry and piece-wise planarity to integrate... | Continue reading
We upgrade [1] to a complete proof of the conjecture NP = PSPACE. [1]: L. Gordeev, E. H. Haeusler, Proof Compression and NP Versus PSPACE,Studia Logica (107) (1): 55-83 (2019) | Continue reading
One major problem with cyberbullying research is the lack of data, sinceresearchers are traditionally forced to rely on survey data where victims andperpetrators self-report their impressions.... | Continue reading
Modern microprocessors are equipped with Single Instruction Multiple Data(SIMD) or vector instructions which expose data level parallelism at a finegranularity. Programmers exploit this... | Continue reading
Decrypting and inspecting encrypted malicious communications may assist crimedetection and prevention. Access to client or server memory enables thediscovery of artefacts required for decrypting... | Continue reading
In the name of meritocracy, modern economies devote increasing amounts ofresources to quantifying and ranking the performance of individuals andorganisations. Rankings send out powerful signals,... | Continue reading
We introduce a formalism that exploits the many-input many-output nature ofnodes in quantum circuits. There is a diagrammatic and an algebraic version,the latter similar to the spinor formalism... | Continue reading
While the role of humans is increasingly recognized in machine learningcommunity, representation of and interaction with models in currenthuman-in-the-loop machine learning (HITL-ML) approaches... | Continue reading
While sequence-to-sequence (seq2seq) models achieve state-of-the-artperformance in many natural language processing tasks, they can be too slow forreal-time applications. One performance... | Continue reading
The vanilla LSTM has become one of the most potential architectures inword-level language modeling, like other recurrent neural networks, overfittingis always a key barrier for its... | Continue reading
This paper introduces a combinatorial optimization approach to registerallocation and instruction scheduling, two central compiler problems.Combinatorial optimization has the potential to solve... | Continue reading
We propose a simple method for post-processing the outputs of a textsummarization system in order to refine its overall quality. Our approach is totrain text-to-text rewriting models to correct... | Continue reading
In previous work we gave a mathematical foundation, referred to as DisCoCat,for how words interact in a sentence in order to produce the meaning of thatsentence. To do so, we exploited the... | Continue reading
A recent research line has obtained strong results on bilingual lexiconinduction by aligning independently trained word embeddings in two languagesand using the resulting cross-lingual... | Continue reading
Open book question answering is a type of natural language based QA (NLQA)where questions are expected to be answered with respect to a given set of openbook facts, and common knowledge about a... | Continue reading
A large number of machine translation approaches have recently been developedto facilitate the fluid migration of content across languages. However, theliterature suggests that many obstacles... | Continue reading
The Winograd Schema Challenge (WSC), proposed by Levesque et al. (2011) as analternative to the Turing Test, was originally designed as a pronoun resolutionproblem that cannot be solved based on... | Continue reading
We introduce an electro-optic hardware platform for nonlinear activationfunctions in optical neural networks. The optical-to-optical nonlinearityoperates by converting a small portion of the... | Continue reading
We proposed the info intervention, which intervening the information sendingout from a node. We point out issues of existing definition of \textit{perfectintervention} and claims that info... | Continue reading
The recent European General Data Protection Regulation (GDPR) and other dataprotection regulations restrict the processing of some categories of personaldata (health, political orientation,... | Continue reading
We present a system for online composite event recognition over streamingpositions of commercial vehicles. Our system employs a data enrichment module,augmenting the mobility data with external... | Continue reading
Asteroid impacts pose a major threat to humanity. The age of the dinosaur wasabruptly ended by a 10-km-diameter asteroid. Currently, a nuclear device is theonly means of deflecting large... | Continue reading
Whereas conventional spoken language understanding (SLU) systems map speechto text, and then text to intent, end-to-end SLU systems map speech directly tointent through a single trainable model.... | Continue reading
The combination of machines and humans for translation is effective, withmany studies showing productivity gains when humans post-editmachine-translated output instead of translating from... | Continue reading
Sentence function is an important linguistic feature referring to a user'spurpose in uttering a specific sentence. The use of sentence function has shownpromising results to improve the... | Continue reading
Direct manipulation interfaces provide intuitive and interactive features toa broad range of users, but they often exhibit two limitations: the built-infeatures cannot possibly cover all use... | Continue reading
In this paper, we present a novel technique for a non-parallel voiceconversion (VC) with the use of cyclic variational autoencoder (CycleVAE)-basedspectral modeling. In a variational... | Continue reading
This document is a short and informal tutorial on some aspects of calculatingphase diagrams with the ATAT-tools emc2 and phb and on creating clusterexpansions with maps. It is neither complete,... | Continue reading
The computations required for deep learning research have been doubling everyfew months, resulting in an estimated 300,000x increase from 2012 to 2018 [2].These computations have a surprisingly... | Continue reading
We present SpanBERT, a pre-training method that is designed to betterrepresent and predict spans of text. Our approach extends BERT by (1) maskingcontiguous random spans, rather than random... | Continue reading
In this paper we consider Epistemic Logic Programs, which extend Answer SetProgramming (ASP) with "epistemic operators" and "epistemic negation", and arecent approach to the semantics of such... | Continue reading
Past research has studied two approaches to utilise predefined policy sets inrepeated interactions: as experts, to dictate our own actions, and as types, tocharacterise the behaviour of other... | Continue reading
Many animals, and an increasing number of artificial agents, displaysophisticated capabilities to perceive and manipulate objects. But human beingsremain distinctive in their capacity for... | Continue reading
Answer Set Programming (ASP) is a logic programming paradigm featuring apurely declarative language with comparatively high modeling capabilities.Indeed, ASP can model problems in NP in a... | Continue reading
Depthwise convolution is becoming increasingly popular in modern efficientConvNets, but its kernel size is often overlooked. In this paper, wesystematically study the impact of different kernel... | Continue reading
A heuristic diagram of the evolution of the standard genetic code ispresented. It incorporates, in a way that resembles the energy levels of anatom, the physical notion of broken symmetry and it... | Continue reading
Abstract Dialectical Frameworks (ADFs) are argumentation frameworks whereeach node is associated with an acceptance condition. This allows us to modeldifferent types of dependencies as supports... | Continue reading
We propose a new model for forming beliefs and learning about unknownprobabilities (such as the probability of picking a red marble from a bag withan unknown distribution of coloured marbles).... | Continue reading
This paper deals with robust optimization applied to network flows. Tworobust variants of the minimum-cost integer flow problem are considered.Thereby, uncertainty in problem formulation is... | Continue reading
In this document we describe a rationale for a research program aimed atbuilding an open ``assistant'' in the game Minecraft, in order to make progresson the problems of natural language... | Continue reading
A large amount of data accommodated in knowledge graphs (KG) is actuallymetric. For example, the Wikidata KG contains a plenitude of metric facts aboutgeographic entities like cities, chemical... | Continue reading
Repeated executions of reasoning tasks for varying inputs are necessary inmany applicative settings, such as stream reasoning. In this context, wepropose an incremental grounding approach for... | Continue reading
Deep learning techniques have become the method of choice for researchersworking on algorithmic aspects of recommender systems. With the stronglyincreased interest in machine learning in... | Continue reading
In this paper we explore the application of methods for classical judgmentaggregation in pooling probabilistic opinions on logically related issues. Forthis reason, we first modify the Boolean... | Continue reading
The key cryptographic protocols used to secure the internet and financialtransactions of today are all susceptible to attack by the development of asufficiently large quantum computer. One... | Continue reading
With the rising societal demand for more information-processing capacity withlower power consumption, alternative architectures inspired by the parallelismand robustness of the human brain have... | Continue reading