Showing 1–12 of 12 results for author: Fenech, K

Rational Gaussian wavelets and corresponding model driven neural networks

Authors: Attila Miklós Ámon, Kristian Fenech, Péter Kovács, Tamás Dózsa

Abstract: In this paper we consider the continuous wavelet transform using Gaussian wavelets multiplied by an appropriate rational term. The zeros and poles of this rational modifier act as free parameters and their choice highly influences the shape of the mother wavelet. This allows the proposed construction to approximate signals with complex morphology using only a few wavelet coefficients. We show that… ▽ More In this paper we consider the continuous wavelet transform using Gaussian wavelets multiplied by an appropriate rational term. The zeros and poles of this rational modifier act as free parameters and their choice highly influences the shape of the mother wavelet. This allows the proposed construction to approximate signals with complex morphology using only a few wavelet coefficients. We show that the proposed rational Gaussian wavelets are admissible and provide numerical approximations of the wavelet coefficients using variable projection operators. In addition, we show how the proposed variable projection based rational Gaussian wavelet transform can be used in neural networks to obtain a highly interpretable feature learning layer. We demonstrate the effectiveness of the proposed scheme through a biomedical application, namely, the detection of ventricular ectopic beats (VEBs) in real ECG measurements. △ Less

Submitted 3 February, 2025; originally announced February 2025.

Comments: Submitted to IEEE Transactions on Signal Processing, 2024 (under review)

MSC Class: 65D15 ACM Class: G.1.2

Skel3D: Skeleton Guided Novel View Synthesis

Authors: Aron Fóthi, Bence Fazekas, Natabara Máté Gyöngyössy, Kristian Fenech

Abstract: In this paper, we present an approach for monocular open-set novel view synthesis (NVS) that leverages object skeletons to guide the underlying diffusion model. Building upon a baseline that utilizes a pre-trained 2D image generator, our method takes advantage of the Objaverse dataset, which includes animated objects with bone structures. By introducing a skeleton guide layer following the existin… ▽ More In this paper, we present an approach for monocular open-set novel view synthesis (NVS) that leverages object skeletons to guide the underlying diffusion model. Building upon a baseline that utilizes a pre-trained 2D image generator, our method takes advantage of the Objaverse dataset, which includes animated objects with bone structures. By introducing a skeleton guide layer following the existing ray conditioning normalization (RCN) layer, our approach enhances pose accuracy and multi-view consistency. The skeleton guide layer provides detailed structural information for the generative model, improving the quality of synthesized views. Experimental results demonstrate that our skeleton-guided method significantly enhances consistency and accuracy across diverse object categories within the Objaverse dataset. Our method outperforms existing state-of-the-art NVS techniques both quantitatively and qualitatively, without relying on explicit 3D representations. △ Less

Submitted 4 December, 2024; originally announced December 2024.

Introducing MeMo: A Multimodal Dataset for Memory Modelling in Multiparty Conversations

Authors: Maria Tsfasman, Bernd Dudzik, Kristian Fenech, Andras Lorincz, Catholijn M. Jonker, Catharine Oertel

Abstract: Conversational memory is the process by which humans encode, retain and retrieve verbal, non-verbal and contextual information from a conversation. Since human memory is selective, differing recollections of the same events can lead to misunderstandings and misalignments within a group. Yet, conversational facilitation systems, aimed at advancing the quality of group interactions, usually focus on… ▽ More Conversational memory is the process by which humans encode, retain and retrieve verbal, non-verbal and contextual information from a conversation. Since human memory is selective, differing recollections of the same events can lead to misunderstandings and misalignments within a group. Yet, conversational facilitation systems, aimed at advancing the quality of group interactions, usually focus on tracking users' states within an individual session, ignoring what remains in each participant's memory after the interaction. Understanding conversational memory can be used as a source of information on the long-term development of social connections within a group. This paper introduces the MeMo corpus, the first conversational dataset annotated with participants' memory retention reports, aimed at facilitating computational modelling of human conversational memory. The MeMo corpus includes 31 hours of small-group discussions on Covid-19, repeated 3 times over the term of 2 weeks. It integrates validated behavioural and perceptual measures, audio, video, and multimodal annotations, offering a valuable resource for studying and modelling conversational memory and group dynamics. By introducing the MeMo corpus, analysing its validity, and demonstrating its usefulness for future research, this paper aims to pave the way for future research in conversational memory modelling for intelligent system development. △ Less

Submitted 15 October, 2024; v1 submitted 7 September, 2024; originally announced September 2024.

Perceived personality state estimation in dyadic and small group interaction with deep learning methods

Authors: Kristian Fenech, Ádám Fodor, Sean P. Bergeron, Rachid R. Saboundji, Catharine Oertel, András Lőrincz

Abstract: Dyadic and small group collaboration is an evolutionary advantageous behaviour and the need for such collaboration is a regular occurrence in day to day life. In this paper we estimate the perceived personality traits of individuals in dyadic and small groups over thin-slices of interaction on four multimodal datasets. We find that our transformer based predictive model performs similarly to human… ▽ More Dyadic and small group collaboration is an evolutionary advantageous behaviour and the need for such collaboration is a regular occurrence in day to day life. In this paper we estimate the perceived personality traits of individuals in dyadic and small groups over thin-slices of interaction on four multimodal datasets. We find that our transformer based predictive model performs similarly to human annotators tasked with predicting the perceived big-five personality traits of participants. Using this model we analyse the estimated perceived personality traits of individuals performing tasks in small groups and dyads. Permutation analysis shows that in the case of small groups undergoing collaborative tasks, the perceived personality of group members clusters, this is also observed for dyads in a collaborative problem solving task, but not in dyads under non-collaborative task settings. Additionally, we find that the group level average perceived personality traits provide a better predictor of group performance than the group level average self-reported personality traits. △ Less

Submitted 9 November, 2022; originally announced November 2022.

Comments: 11 pages, 4 figures

Minimizing false negative rate in melanoma detection and providing insight into the causes of classification

Authors: Ellák Somfai, Benjámin Baffy, Kristian Fenech, Changlu Guo, Rita Hosszú, Dorina Korózs, Fabrizio Nunnari, Marcell Pólik, Daniel Sonntag, Attila Ulbert, András Lőrincz

Abstract: Our goal is to bridge human and machine intelligence in melanoma detection. We develop a classification system exploiting a combination of visual pre-processing, deep learning, and ensembling for providing explanations to experts and to minimize false negative rate while maintaining high accuracy in melanoma detection. Source images are first automatically segmented using a U-net CNN. The result o… ▽ More Our goal is to bridge human and machine intelligence in melanoma detection. We develop a classification system exploiting a combination of visual pre-processing, deep learning, and ensembling for providing explanations to experts and to minimize false negative rate while maintaining high accuracy in melanoma detection. Source images are first automatically segmented using a U-net CNN. The result of the segmentation is then used to extract image sub-areas and specific parameters relevant in human evaluation, namely center, border, and asymmetry measures. These data are then processed by tailored neural networks which include structure searching algorithms. Partial results are then ensembled by a committee machine. Our evaluation on the largest skin lesion dataset which is publicly available today, ISIC-2019, shows improvement in all evaluated metrics over a baseline using the original images only. We also showed that indicative scores computed by the feature classifiers can provide useful insight into the various features on which the decision can be based. △ Less

Submitted 9 March, 2021; v1 submitted 18 February, 2021; originally announced February 2021.

Comments: supplementary materials included

ACM Class: I.4.9; J.3

Isolation of $k$-cliques

Authors: Peter Borg, Kurt Fenech, Pawaton Kaemawichanurat

Abstract: For any positive integer $k$ and any $n$-vertex graph $G$, let $ι(G,k)$ denote the size of a smallest set $D$ of vertices of $G$ such that the graph obtained from $G$ by deleting the closed neighbourhood of $D$ contains no $k$-clique. Thus, $ι(G,1)$ is the domination number of $G$. We prove that if $G$ is connected, then $ι(G,k) \leq \frac{n}{k+1}$ unless $G$ is a $k$-clique or $k = 2$ and $G$ is… ▽ More For any positive integer $k$ and any $n$-vertex graph $G$, let $ι(G,k)$ denote the size of a smallest set $D$ of vertices of $G$ such that the graph obtained from $G$ by deleting the closed neighbourhood of $D$ contains no $k$-clique. Thus, $ι(G,1)$ is the domination number of $G$. We prove that if $G$ is connected, then $ι(G,k) \leq \frac{n}{k+1}$ unless $G$ is a $k$-clique or $k = 2$ and $G$ is a $5$-cycle. The bound is sharp. The case $k=1$ is a classical result of Ore, and the case $k=2$ is a recent result of Caro and Hansberg. Our result solves a problem of Caro and Hansberg. △ Less

Submitted 28 December, 2018; originally announced December 2018.

Comments: 7 pages

MSC Class: 05D05; 05C35; 05C69

Irregular independence and irregular domination

Authors: Peter Borg, Yair Caro, Kurt Fenech

Abstract: If $A$ is an independent set of a graph $G$ such that the vertices in $A$ have different degrees, then we call $A$ an irregular independent set of $G$. If $D$ is a dominating set of $G$ such that the vertices that are not in $D$ have different numbers of neighbours in $D$, then we call $D$ an irregular dominating set of $G$. The size of a largest irregular independent set of $G$ and the size of a… ▽ More If $A$ is an independent set of a graph $G$ such that the vertices in $A$ have different degrees, then we call $A$ an irregular independent set of $G$. If $D$ is a dominating set of $G$ such that the vertices that are not in $D$ have different numbers of neighbours in $D$, then we call $D$ an irregular dominating set of $G$. The size of a largest irregular independent set of $G$ and the size of a smallest irregular dominating set of $G$ are denoted by $α_{ir}(G)$ and $γ_{ir}(G)$, respectively. We initiate the investigation of these two graph parameters. For each of them, we obtain sharp bounds in terms of basic graph parameters such as the order, the size, the minimum degree and the maximum degree, and we obtain Nordhaus-Gaddum-type bounds. We also establish sharp bounds relating the two parameters. Furthermore, we characterize the graphs $G$ with $α_{ir}(G)=1$, we determine those that are planar, and we determine those that are outerplanar. △ Less

Submitted 21 June, 2017; originally announced June 2017.

Comments: 15 pages

MSC Class: 05C69; 05C35; 05C10

Comparison of strong-coupling theories for a two-dimensional Fermi gas

Authors: Brendan C. Mulkerin, Kristian Fenech, Paul Dyke, Chris J. Vale, Xia-Ji Liu, Hui Hu

Abstract: Understanding the formation of Cooper pairs and the resulting thermodynamic properties of a low-dimensional Fermi gas is an important area of research, elucidating our understanding of high temperature superconductors. In lower dimensions quantum fluctuations are expected to play an increasingly important role and the reliability of strong-coupling theories becomes questionable. Here, we present a… ▽ More Understanding the formation of Cooper pairs and the resulting thermodynamic properties of a low-dimensional Fermi gas is an important area of research, elucidating our understanding of high temperature superconductors. In lower dimensions quantum fluctuations are expected to play an increasingly important role and the reliability of strong-coupling theories becomes questionable. Here, we present a comparison of recent thermodynamic measurements and theoretical predictions from different many-body $T$-matrix theories for a two-dimensional strongly interacting Fermi gas in the normal state. We find that the fully self-consistent $T$-matrix theory provides the best description of the experimental data over a wide range of temperatures and interatomic interactions. Our comparison reveals the crucial role played by the interactions between Cooper pairs and suggests that the future development of a quantitative strong-coupling theory for two-dimensional Fermi superfluids must explicitly take into account the diagrams that are responsible for pair-pair interactions. △ Less

Submitted 15 February, 2016; v1 submitted 28 September, 2015; originally announced September 2015.

Comments: Five pages, four figures

Journal ref: Phys. Rev. A 92, 063636 (2015)

Thermodynamics of an attractive 2D Fermi gas

Authors: K. Fenech, P. Dyke, T. Peppler, M. G. Lingham, S. Hoinka, H. Hu, C. J. Vale

Abstract: Thermodynamic properties of matter are conveniently expressed as functional relations between variables known as equations of state. Here we experimentally determine the compressibility, density and pressure equations of state for an attractive 2D Fermi gas in the normal phase as a function of temperature and interaction strength. In 2D, interacting gases exhibit qualitatively different features t… ▽ More Thermodynamic properties of matter are conveniently expressed as functional relations between variables known as equations of state. Here we experimentally determine the compressibility, density and pressure equations of state for an attractive 2D Fermi gas in the normal phase as a function of temperature and interaction strength. In 2D, interacting gases exhibit qualitatively different features to those found in 3D. This is evident in the normalized density equation of state, which peaks at intermediate densities corresponding to the crossover from classical to quantum behaviour. △ Less

Submitted 10 February, 2016; v1 submitted 18 August, 2015; originally announced August 2015.

Comments: Contains minor revisions

Journal ref: Phys. Rev. Lett. 116, 045302 (2016)

Criteria for 2D kinematics in an interacting Fermi gas

Authors: P. Dyke, K. Fenech, T. Peppler, M. G. Lingham, S. Hoinka, W. Zhang, B. Mulkerin, H. Hu, X. -J. Liu, C. J. Vale

Abstract: Ultracold Fermi gases subject to tight transverse confinement offer a highly controllable setting to study the two-dimensional (2D) BCS to Berezinskii-Kosterlitz-Thouless superfluid crossover. Achieving the 2D regime requires confining particles to their transverse ground state which presents challenges in interacting systems. Here, we establish the conditions for an interacting Fermi gas to behav… ▽ More Ultracold Fermi gases subject to tight transverse confinement offer a highly controllable setting to study the two-dimensional (2D) BCS to Berezinskii-Kosterlitz-Thouless superfluid crossover. Achieving the 2D regime requires confining particles to their transverse ground state which presents challenges in interacting systems. Here, we establish the conditions for an interacting Fermi gas to behave kinematically 2D. Transverse excitations are detected by measuring the transverse expansion rate which displays a sudden increase when the atom number exceeds a critical value $N_{2D}$ signifying a density driven departure from 2D kinematics. For weak interactions $N_{2D}$ is set by the aspect ratio of the trap. Close to a Feshbach resonance, however, the stronger interactions reduce $N_{2D}$ and excitations appear at lower density. △ Less

Submitted 14 January, 2016; v1 submitted 17 November, 2014; originally announced November 2014.

Comments: Replaced with published version, includes supplementary information

Journal ref: Phys. Rev. A 93, 011603 (2016)

Local observation of pair-condensation in a Fermi gas at unitarity

Authors: M. G. Lingham, K. Fenech, S. Hoinka, C. J. Vale

Abstract: We present measurements of the local (homogeneous) density-density response function of a Fermi gas at unitarity using spatially resolved Bragg spectroscopy. By analyzing the Bragg response across one axis of the cloud we extract the response function for a uniform gas which shows a clear signature of the Bose-Einstein condensation of pairs of fermions when the local temperature drops below the su… ▽ More We present measurements of the local (homogeneous) density-density response function of a Fermi gas at unitarity using spatially resolved Bragg spectroscopy. By analyzing the Bragg response across one axis of the cloud we extract the response function for a uniform gas which shows a clear signature of the Bose-Einstein condensation of pairs of fermions when the local temperature drops below the superfluid transition temperature. The method we use for local measurement generalizes a scheme for obtaining the local pressure in a harmonically trapped cloud from the line density and can be adapted to provide any homogeneous parameter satisfying the local density approximation. △ Less

Submitted 19 March, 2014; v1 submitted 26 August, 2013; originally announced August 2013.

Comments: New version, 6 pages, 3 figures

Journal ref: Phys. Rev. Lett. 112, 100404 (2014)

Precise determination of the structure factor and contact in a unitary Fermi gas

Authors: Sascha Hoinka, Marcus Lingham, Kristian Fenech, Hui Hu, Chris J. Vale, Joaquín E. Drut, Stefano Gandolfi

Abstract: We present a high-precision determination of the universal contact parameter in a strongly interacting Fermi gas. In a trapped gas at unitarity we find the contact to be $3.06 \pm 0.08$ at a temperature of 0.08 of the Fermi temperature in a harmonic trap. The contact governs the high-momentum (short-range) properties of these systems and this low temperature measurement provides a new benchmark fo… ▽ More We present a high-precision determination of the universal contact parameter in a strongly interacting Fermi gas. In a trapped gas at unitarity we find the contact to be $3.06 \pm 0.08$ at a temperature of 0.08 of the Fermi temperature in a harmonic trap. The contact governs the high-momentum (short-range) properties of these systems and this low temperature measurement provides a new benchmark for the zero temperature homogeneous contact. The experimental measurement utilises Bragg spectroscopy to obtain the dynamic and static structure factors of ultracold Fermi gases at high momentum in the unitarity and molecular Bose-Einstein condensate (BEC) regimes. We have also performed quantum Monte Carlo calculations of the static properties, extending from the weakly coupled Bardeen-Cooper-Schrieffer (BCS) regime to the strongly coupled BEC case, which show agreement with experiment at the level of a few percent. △ Less

Submitted 4 February, 2013; v1 submitted 17 September, 2012; originally announced September 2012.

Comments: Replaced with accepted version

Journal ref: Phys. Rev. Lett. 110, 055305 (2013)