Transfer Learning for Human & AI

The Post Correspondence Problem is a classical decision problem about equalisers of free monoid homomorphisms . We prove connections between severalvariations of this classical problem, but in the setting of free groups and free group homomorphism . Among other results, and working under certaininjectivity assumptions, we prove that computing the rank of the equaliser of apair of free grouphomomorphisms can be applied to computing a basis of this equaliser .…

Deep Neural Networks are known to be vulnerable to small, adversariallycrafted, perturbations . The current most effective defense methods against these adversarial attacks are variants of adversarial training . In this paper, we introduce a radically different defense trained only on clean images .…

The current state-of-the-art model HiAGM for hierarchical text classification has two limitations . We propose HTCInfoMax to addressthese issues by introducing information maximization which includes twomodules . The first module can model interaction between each text sample and its ground truth labels explicitly which filters out irrelevant information .…

We consider the problem of ON-OFF privacy in which a user is interested inthe latest message generated by one of n sources available at a server . The challenge of allowing the privacy tobe toggled between ON and OFF is that the user’s online behavior is correlated over time .…

We treat arrays aseagerly-memoized functions on typed index sets, allowing abstract functionmanipulations, such as currying, to work on arrays . We argue for an explicit nested indexing style that mirrors application of functions to arguments . We also introduce afine-grained typed effects system which affords concise andautomatically-parallelized in-place updates .…

Cloning is a general approach to create new functionality within variants as well as new system variants . It is a fast, flexible, intuitive, and economical approach to evolve systems in the short run . However, clones within a variant are usually not considered in the process, but are also a cause for unsustainable software .…

Mobility and transport involve crowds and require multiple stakeholders – such as policy-makers, planners, transport operators, and the travelers themselves . Traditional approaches have been focused on time savings, proposing to users solutions that include the shortest or fastest paths .…

We present a SE(3)-equivariant graph neural network (GNN) approach that predicts the formation factor and effective permeability from micro-CT images . FFT solvers are established to compute both the formationfactor and effectiveness permeability, while the topology and geometry of the porespace are represented by a persistence-based Morse graph .…

In mobile games industry, the churn rate is often pronounced due to the high competition and cost in customer acquisition . This needs churn prediction, predicting users who will be churning within a given time period . Accuratechurn prediction can enable the businesses to devise and engage strategicremediations to maintain a low churn rate.…

We propose to use speech cues (on what, when and how the user talks) as an objective real-time measure of trust . This could be implemented in the robot to calibrate towards appropriate trust . However, we would like to open the discussion on how to deal with the ethical implications of this trust measure .…

Automatic garbage collection (GC) prevents certain kinds of bugs and reducesprogramming overhead . GC techniques for sequential programs are based onreachability analysis . However, testing reachability from a root set is inadequate for determining whether an actor is garbage . We present a low-overhead reference listing technique (called DRL) fortermination detection in actor systems .…

Anomaly detection is designed to profile thenormal behavior of computer programs in order to detect intrusions and errors as anomalous deviations from the observed normal . Normal butunobserved behavior can trigger false positives . This limitation has significantly decreased the practical viability of anomaly detectiontechniques .…

Symmetries have proven to be important ingredients in the analysis of neuralnetworks . We undertake a systematic study of the role that symmetry plays . Symmetry alone does not specify the optimization path, but the more symmetries are contained in the model the more restrictions are imposed on the path .…

This paper presents a Matlab toolbox to perform basic image processing and visualization tasks, particularly designed for medical image processing . The toolbox is entirely written in native Matlab code, but is fast and flexible . Main use cases for the toolbox are illustrated here, including image input/output, pre-processing, filtering, image registration and visualisation .…

Collaborative Mobile Crowdsourcing (CMC) enables task requesters tohire groups of IoT devices’ users that must communicate with each other and coordinate their operational activities in order to accomplish complex tasks . The vast majority of previous studies focused on non-cooperative MC schemes in Internet of Things (IoT) systems .…

This paper presents a biologically plausible method for converting real-valued input into spike trains for processing with spiking neuralnetworks . The proposed method mimics the adaptive behaviour of retinal ganglioncells and allows input neurons to adapt their response to changes in thestatistics of the input .…

Fixed low-resolution Analog to Digital Converters (ADC) help reduce power consumption in millimeter-wave Massive Multiple Multiple-Input Multiple-Output (Ma-MIMO) receivers . However, they do not guarantee optimal energy efficiency (EE) Adopting variable-resolution ADCs in Ma-MimO receivers can improve performance with Mean Squared Error(MSE) and throughput while providing better EE .…

MCAS (Memory Centric Active Storage) is a persistent memory tier for durable data storage . It is designed from the ground-up toprovide a key-value capability with low-latency guarantees and data durability . To reduce data movement and makefurther gains in performance, we provide support for user-defined “push-down”operations (known as Active Data Objects) that can execute directly and safely on the value-memory associated with one or more keys .…

Recent research shows that Deep Neuroevolution (DNE) is alsocapable of evolving multi-million-parameter DNN’s . So far, DNE has only been applied to complex single-agent problems . In this paper, we describe and validate a new approachbased on Coevolution . To validate our approach, we benchmark two DeepCoevolutionary Algorithms on a range of multi-agent Atari games and compare our results against the results of Ape-X DQN .…

Deep learning methods have reached state-of-the-art performance in cardiac image segmentation . Currently, the main bottleneck towards their effective translation into clinics requires assuring continuous high model performance . In this work, we present a novel learning framework to monitor the performance of heart segmentation models in the absence of ground truth .…

In this paper, we study how to efficiently and reliably detect active devicesand estimate their channels in a multiple- input multiple-input multiple-output (MIMO)orthogonal frequency-division multiplexing (OFDM) system . We propose a block-wise linear channel model and an efficient turbo message passing (Turbo-MP) algorithm to resolve the Bayesian inference problem with near-linearcomplexity .…

Robustness and counterfactual bias are usually evaluated on a test dataset . If the test dataset is perturbedslightly, will the evaluation results keep the same? In this paper, we propose a “double perturbation” framework to uncover model weaknesses beyond the test data .…

We propose a novel approach for the adaptivechoice of the learning rate in first-order gradient descent methods . Our approach modifies the objective function by adding a regularization term on thelearning rate . We theoreticallyprove that some well-designed regularizers can improve the convergenceperformance under the strong-convexity condition of the objectivefunction .…

We propose a unitary precoding scheme to improve polar-coded MIMO systems . The proposed polar-precoding scheme relies on the polarization criterion . Our simulation results show that the proposal outperforms the state-of-the-art DFT pre-code scheme . In contrast to the traditional design of MIMo precoding criteria, we design a matrix for maximizing polarization among data streams without degrading the capacity .…

Phase-Amplitude Coupling is a type of Cross-Frequency Coupling (CFC) method, consisting in measure the synchronization of phase and amplitude for the different EEG bands andelectrodes . This provides information regarding brain areas that aresynchronously activated, and eventually a marker of functional connectivity between these areas .…

In domains where users tend to develop long-term preferences that do notchange too frequently, stability of recommendations is an important factor of the perceived quality of a recommender system . We propose an incrementallearning scheme that mitigates such problems through the dynamic modeling approach .…

There are many potential measures that one might use when evaluating mixed reality experiences . In this position paper I will argue that there arevarious stances to take for evaluation, depending on the framing of the work . I will sketch out some directions for developing more robust measures that can help the field move forward.…

The genome encoding and the crossover operator are extended to make them applicable to layered networks . The proposed layout enables the size and shape of individual kernels within a convolutional layer to be evolved with acorresponding new mutation operator. The proposed framework employs a hybridoptimisation strategy involving structural changes through epigenetic evolutionand weight update through backpropagation in a population-based setting .…

Social VR enables people to interact over distance with others in real-time . Remote people, typically represented as avatars, to communicate andperform activities together in a join shared virtual environment . End-users perceived the photo-realistic experience of SocialVRsimilar to face-to-face scenarios and appreciated this new creative medium .…

We establish theoretical relationship between the level of balance in a network and stockpredictability . We use rank correlations as distance functions to establish theinterconnectivity between stock returns . We find a clear balance-unbalance transition for six of the nine countries, following the August 2011 Black Monday in the US, when the Economic PolicyUncertainty index for this country reached its highest monthly level before theCOVID-19 crisis .…

In the Approval Participatory Budgeting problem an agent prefers a set of projects $W’$ over $W$ if she approves strictly more projects . The core is nonempty if there are no other sets of projects that both prefer $WW$ and $K$ can fund it .…

In this paper, we propose an end-to-end Mandarin tone classification method from using both the spectrogram and the shortterm context information as the inputs . We first divide thespectrogram frames into syllable segments using force alignment results produced by an ASR model .…

GenerativeAdversarial Networks (GANs) involve training of overparameterized models where the number of parameters of the model exceeds a certain threshold . The role and benefits of modeloverparameterization in the convergence of GDA to a global saddle point in non-convex concave min-max problems is far less understood .…

We study how to introduce locality mechanisms into vision transformers . Locally-enhanced transformers outperform the baselines DeiT-T and PVT-t by 2.6\% and 3.1\% with a negligible increase in the number of parameters and effort . The same mechanism was used to apply to 4 visiontransformers, which shows the generalization of the locality concept .…

Blockchains have witnessed widespread adoption in the past decade in various fields . The growing demand makes their scalability and sustainability challenges more evident than ever . More and more blockchains have begun to adopt proof-of-stake (PoS) consensus protocols to address those challenges .…

The bundle recommendation model has been deployed in production formore than one year in a popular online game developed by Netease Games . The model yields more than 60\% improvement on conversion rate of Bundles, and a relative improvement of more than 15\% on gross merchandisevolume .…

Causal Discovery methods aim to identify a DAG structure that represents causal relationships from observational data . We stress that it is important to test such methods for robustness in practical settings . We conclude that NOTEARS is not suitable for identifying truly causalrelationships from the data .…

A more intense collaboration can result into increased productivity which can potentially have an impact on the author’s future$h$-index . The proposed method outperforms standard machinelearning techniques based on simple graph metrics along with noderepresentations learned from the textual content of the authors’ papers .…

Integration, just as much as differentiation, is a fundamental calculus tool that is widely used in many scientific domains . Formalizing the mathematicalconcept of integration and the associated results in a formal proof assistant helps providing the highest confidence on the correction of numerical programs involving the use of integration, directly or indirectly .…

Spiking neural networks have shown great promise for the design of low-powersensory-processing and edge-computing hardware platforms . However, implementing on-chip learning algorithms on such architectures is still an open challenge, especially for multi-layer networks that rely on the back-propagational algorithm .…

Pre-trained language models such as BERT have become a more common choice of natural language processing (NLP) tasks . We find that the word vectors are not centered around the origin, and the average cosine similarity betweentwo random words is much higher than zero .…

Hypergraphs are data structures capable of capturing supra-dyadic relations . We can use them to model binary relations, but also to model groups of groups of entities . As more information is added to the hypergraph, it also becomes denser, making the task of ranking nodes or hyperedges more complex .…

The problem of the real-time multiple-input multiple- input multiple-output (MIMO) arraycontrol is addressed by means of an innovative hybrid beamforming technique . The synthesis of the excitationsof the MIMO system is first re-formulated as that of matching an ideal “hybrid” pattern fitting capacity or low-interference constraints along the angularcoordinates .…

Terahertz (THz) communication is gaining more interest as one of the enablers of high-data-rate short-distance indoor applications inbeyond 5G networks . Non-orthogonal multiple-access (NOMA)-enabled schemes are promising schemes to realize the target spectral efficiency, lowlatency, and user fairness requirements in future networks .…

Entropoid Diffie-Hellman problem is hard in Sylow $q$-subquasigroups . We post a conjecture that DEDHP is hard for Sylow- $q$.-subaquasig Groups . The entropoid based cryptographic primitives are supposed to be resistant to quantum algorithms . We give a proof-of-concept implementation in SageMath 9.2 for all proposed algorithms and schemes in an appendix .…

We consider the problem of pure exploration with subset-wise preferencefeedback, which contains $N$ arms with features . The learner is allowed to query subsets of size $K$ and receives feedback in the form of a noisy winner . This setting is relevant in various online decision-makingscenarios involving human feedback such as online retailing, streamingservices, news feed, and online advertising; since it is easier and more reliable for people to choose a preferred item from a subset than to assign alikability score to an item in isolation .…

Reference-free evaluation has the potential to make machine translation evaluation substantially more scalable, allowing us to pivot easily to new languages or domains . It has been recently shown that the probabilities givenby a large, multilingual model can achieve state of the art results when used as a reference-free metric .…

Complex decision making problems such as the privacy policy selection whensharing content in online social networks can significantly benefit from artificial intelligence systems . With the use of Computational Argumentation,it is possible to persuade human users to modify their initial decisions to avoid potential privacy threats and violations .…

We present a novel game-theoretic formulation of Non-Negative MatrixFactorization (NNMF) The NNFF is a popular data-analysis method with many scientific and engineering applications . The NNFF is shown to havefavorable scaling and parallelization properties .…

Flash-based disk cache has been widely used in indurstry in the last decade, such as flashcache (facebook, 2010), bcache (opensource, 2013) These cachesystems typically use commercial SSDs to store and manage cache data . WLFC can significantly reduce the write latency by 70%, increase the write-throughput up to 4.5x, and remove 88.9% unnecessary erase operations at most .…