Dr Alexia Zoumpoulaki
(Mae hi'n)
Staff academaidd ac ymchwil
Yr Ysgol Cyfrifiadureg a Gwybodeg
- Ar gael fel goruchwyliwr ôl-raddedig
Trosolwyg
Rwy'n Ddarlithydd (athro cynorthwyol) mewn Cyfrifiadureg a Gwybodeg ym Mhrifysgol Caerdydd, y DU, sy'n arbenigo mewn AI, dysgu peirianyddol ar gyfer Ceisiadau Ffactorau Dynol. Yn y gorffennol rwyf wedi gweithio mewn amrywiol brosiectau ymchwil rhyngddisgyblaethol, gan gynnwys dosbarthu cof, canfod twyll ac efelychiadau torfol. Mae llawer o'm gwaith wedi canolbwyntio ar ddatblygu dulliau ar gyfer dadansoddi a dosbarthu data cyfres amser.
Ar hyn o bryd mae gen i ddiddordeb mewn prosesu fideos amlfoddol. Mynegiant yr wyneb (micro) - cydnabyddiaeth emosiwn, dadansoddi testun, dadansoddi llais ym maes canfod twyll gan ddefnyddio dysgu peiriant. Mae gen i ddiddordeb mewn ymgorffori nodweddion a dynnwyd o'r dulliau hyn gyda ffactorau seicolegol eraill yn ogystal â nodweddion o olrhain llygaid.
Yn fras, mae gen i ddiddordeb mewn gwyddoniaeth gynhwysfawr (cyfrifiadurol) (niwro) a sut y gall cyfrifiadureg helpu ymddygiad enghreifftiol a phrosesau gwybyddol. Bydd y ddealltwriaeth hon yn ein helpu i adeiladu ceisiadau newydd cyffrous.
Cyhoeddiad
2021
- Finnegan, D., Zoumpoulaki, A. and Eslambolchilar, P. 2021. Does mixed reality have a Cassandra Complex?. Frontiers in Virtual Reality 2, article number: 673547. (10.3389/frvir.2021.673547)
2020
- Bowman, H., Brooks, J. L., Hajilou, O., Zoumpoulaki, A. and Litvak, V. 2020. Breaking the circularity in circul aranalyses: simulations and formal treatment of the flattened average approach. PLoS Computational Biology 16(11), article number: e1008286. (10.1371/journal.pcbi.1008286)
2019
- Alsufyani, A. et al. 2019. Breakthrough percepts of famous faces. Psychophysiology 56(1), article number: e13279. (10.1111/psyp.13279)
2018
- Belal, S. et al. 2018. Identification of memory reactivation during sleep by EEG classification. NeuroImage 176, pp. 203-214. (10.1016/j.neuroimage.2018.04.029)
2017
- Brooks, J. L., Zoumpoulaki, A. and Bowman, H. 2017. Data-driven region-of-interest selection without inflating Type I error rate. Psychophysiology 54(1), pp. 100-113. (10.1111/psyp.12682)
2016
- Zoumpoulaki, A., Gootjes-Dreesbach, L., Bergstrom, Z., Alsufyani, A. and Bowman, H. 2016. Context matters: Driving perceptual breakthrough through contextual priming. Journal of Vision 16(12), article number: 1037. (10.1167/16.12.1037)
2015
- Zoumpoulaki, A., Alsufyani, A., Filetti, M., Brammer, M. and Bowman, H. 2015. Latency as a region contrast: Measuring ERP latency differences with Dynamic Time Warping. Psychophysiology 52(12), pp. 1559-1576. (10.1111/psyp.12521)
- Zoumpoulaki, A., Alsufyani, A. and Bowman, H. 2015. Corrigendum to resampling the peak, some dos and don'ts. Psychophysiology 52(5), pp. 726-726. (10.1111/psyp.12444)
- Zoumpoulaki, A., Alsufyani, A. and Bowman, H. 2015. Resampling the peak, some dos and don'ts. Psychophysiology 52(3), pp. 444-448. (10.1111/psyp.12363)
2013
- Zoumpoulaki, A., Alsufyani, A., Filetti, M., Brammer, M. and Bowman, H. 2013. ERP latency contrasts using Dynamic Time Warping algorithm. BMC Neuroscience 14(Supp 1), pp. P434. (10.1186/1471-2202-14-S1-P434)
- Alsufyani, A., Zoumpoulaki, A., Filetti, M. and Bowman, H. 2013. A new method for detecting deception in Event Related Potentials using individual-specific weight templates [Abstract]. BMC Neuroscience 14(S1), pp. P34. (10.1186/1471-2202-14-S1-P34)
2010
- Zoumpoulaki, A., Avradinis, N. and Vosinakis, S. 2010. A multi-agent simulation framework for emergency evacuations incorporating personality and emotions. Presented at: 6th Hellenic Conference on Artificial Intelligence (SETN 2010), Athens, Greece, 4-7 May 2010 Presented at Konstantopoulos, S. et al. eds.Advances in Artificial Intelligence: Theories, Models, and Applications, Vol. 6040. Lecture Notes in Artificial Intelligence Springer pp. 423-428., (10.1007/978-3-642-12842-4_54)
Cynadleddau
- Zoumpoulaki, A., Avradinis, N. and Vosinakis, S. 2010. A multi-agent simulation framework for emergency evacuations incorporating personality and emotions. Presented at: 6th Hellenic Conference on Artificial Intelligence (SETN 2010), Athens, Greece, 4-7 May 2010 Presented at Konstantopoulos, S. et al. eds.Advances in Artificial Intelligence: Theories, Models, and Applications, Vol. 6040. Lecture Notes in Artificial Intelligence Springer pp. 423-428., (10.1007/978-3-642-12842-4_54)
Erthyglau
- Finnegan, D., Zoumpoulaki, A. and Eslambolchilar, P. 2021. Does mixed reality have a Cassandra Complex?. Frontiers in Virtual Reality 2, article number: 673547. (10.3389/frvir.2021.673547)
- Bowman, H., Brooks, J. L., Hajilou, O., Zoumpoulaki, A. and Litvak, V. 2020. Breaking the circularity in circul aranalyses: simulations and formal treatment of the flattened average approach. PLoS Computational Biology 16(11), article number: e1008286. (10.1371/journal.pcbi.1008286)
- Alsufyani, A. et al. 2019. Breakthrough percepts of famous faces. Psychophysiology 56(1), article number: e13279. (10.1111/psyp.13279)
- Belal, S. et al. 2018. Identification of memory reactivation during sleep by EEG classification. NeuroImage 176, pp. 203-214. (10.1016/j.neuroimage.2018.04.029)
- Brooks, J. L., Zoumpoulaki, A. and Bowman, H. 2017. Data-driven region-of-interest selection without inflating Type I error rate. Psychophysiology 54(1), pp. 100-113. (10.1111/psyp.12682)
- Zoumpoulaki, A., Gootjes-Dreesbach, L., Bergstrom, Z., Alsufyani, A. and Bowman, H. 2016. Context matters: Driving perceptual breakthrough through contextual priming. Journal of Vision 16(12), article number: 1037. (10.1167/16.12.1037)
- Zoumpoulaki, A., Alsufyani, A., Filetti, M., Brammer, M. and Bowman, H. 2015. Latency as a region contrast: Measuring ERP latency differences with Dynamic Time Warping. Psychophysiology 52(12), pp. 1559-1576. (10.1111/psyp.12521)
- Zoumpoulaki, A., Alsufyani, A. and Bowman, H. 2015. Corrigendum to resampling the peak, some dos and don'ts. Psychophysiology 52(5), pp. 726-726. (10.1111/psyp.12444)
- Zoumpoulaki, A., Alsufyani, A. and Bowman, H. 2015. Resampling the peak, some dos and don'ts. Psychophysiology 52(3), pp. 444-448. (10.1111/psyp.12363)
- Zoumpoulaki, A., Alsufyani, A., Filetti, M., Brammer, M. and Bowman, H. 2013. ERP latency contrasts using Dynamic Time Warping algorithm. BMC Neuroscience 14(Supp 1), pp. P434. (10.1186/1471-2202-14-S1-P434)
- Alsufyani, A., Zoumpoulaki, A., Filetti, M. and Bowman, H. 2013. A new method for detecting deception in Event Related Potentials using individual-specific weight templates [Abstract]. BMC Neuroscience 14(S1), pp. P34. (10.1186/1471-2202-14-S1-P34)
- Brooks, J. L., Zoumpoulaki, A. and Bowman, H. 2017. Data-driven region-of-interest selection without inflating Type I error rate. Psychophysiology 54(1), pp. 100-113. (10.1111/psyp.12682)
- Zoumpoulaki, A., Gootjes-Dreesbach, L., Bergstrom, Z., Alsufyani, A. and Bowman, H. 2016. Context matters: Driving perceptual breakthrough through contextual priming. Journal of Vision 16(12), article number: 1037. (10.1167/16.12.1037)
- Zoumpoulaki, A., Alsufyani, A., Filetti, M., Brammer, M. and Bowman, H. 2015. Latency as a region contrast: Measuring ERP latency differences with Dynamic Time Warping. Psychophysiology 52(12), pp. 1559-1576. (10.1111/psyp.12521)
- Zoumpoulaki, A., Alsufyani, A. and Bowman, H. 2015. Resampling the peak, some dos and don'ts. Psychophysiology 52(3), pp. 444-448. (10.1111/psyp.12363)
- Zoumpoulaki, A., Alsufyani, A., Filetti, M., Brammer, M. and Bowman, H. 2013. ERP latency contrasts using Dynamic Time Warping algorithm. BMC Neuroscience 14(Supp 1), pp. P434. (10.1186/1471-2202-14-S1-P434)
- Alsufyani, A., Zoumpoulaki, A., Filetti, M. and Bowman, H. 2013. A new method for detecting deception in Event Related Potentials using individual-specific weight templates [Abstract]. BMC Neuroscience 14(S1), pp. P34. (10.1186/1471-2202-14-S1-P34)
- Zoumpoulaki, A., Avradinis, N. and Vosinakis, S. 2010. A multi-agent simulation framework for emergency evacuations incorporating personality and emotions. Presented at: 6th Hellenic Conference on Artificial Intelligence (SETN 2010), Athens, Greece, 4-7 May 2010 Presented at Konstantopoulos, S. et al. eds.Advances in Artificial Intelligence: Theories, Models, and Applications, Vol. 6040. Lecture Notes in Artificial Intelligence Springer pp. 423-428., (10.1007/978-3-642-12842-4_54)
Ymchwil
Diddordebau ymchwil
Mae gen i ddiddordeb mewn cymhwyso AI, technegau dysgu peirianyddol i adeiladu cymwyseddau ffactorau dynol, gan ganolbwyntio ar feysydd fel sylw, amldasgio a gwneud penderfyniadau. Rwy'n defnyddio mesuriadau ffisiolegol fel syllu, dilation disgyblion, eeg ar y cyd â fideo, testun a dadansoddiad llais i astudio'r meysydd uchod.
Mae gen i ddiddordeb mewn ymddygiadau o dan sefyllfaoedd llawn straen, a sut i ddatblygu technolegau i helpu pobl i wneud penderfyniadau a pherfomance tasgau.
Mae gen i arbenigedd mewn dadansoddi a dosbarthu data niwroddelweddu/cyfres amser.
Addysgu
Meddwl Cyfrifiannol CM6114 / CM6614: Nod y cwrs yw galluogi myfyrwyr i drosi problemau yn y byd go iawn yn god cyfrifiadur. Mae'r ffocws ar ddatblygu'r gallu i gymhwyso cysyniadau craidd trwy ysgrifennu cod sy'n datrys problemau sylfaenol, gan osod y sylfaen ar gyfer codio cynhyrchiol mewn modiwlau diweddarach. Ar ôl cwblhau'r modiwl hwn yn llwyddiannus, disgwylir i fyfyrwyr fynd i'r afael â phroblemau mewn ffordd gyfrifiadurol trwy feddwl rhesymegol, diwygio, tynnu sylw, dadelfennu a chynrychiolaeth ddata briodol. Bydd myfyrwyr hefyd yn datblygu dealltwriaeth sylfaenol o'r algorithmau datrys problemau mwyaf poblogaidd (chwilio, didoli, llwybro, dyrannu adnoddau), eu disgrifio mewn ffordd wyddonol a gwerthuso eu cymhlethdod. Yn olaf, byddant yn ymdrin â sut mae gwybodaeth yn cael ei storio a'i chyrchu mewn cyfrifiadur a byddant yn gallu cyfieithu ar draws gwahanol systemau rhifau.
Goruchwyliaeth prosiect israddedig ac ôl-raddedig: Mae'r prosiectau presennol a'r gorffennol yn cynnwys:
- Canfod Osgiliad Amledd Uchel Gan ddefnyddio Dadansoddiad Wavelet a Rhwydweithiau Niwclear Cyfnewidiol
- Canfod twyll mewn testun gyda Dysgu Tranfer
- Cais ar gyfer cymharu a delweddu piblinellau cyn-brosesu M/EEG
- Canfod Celwyddau gan ddefnyddio Eyetracker a dadansoddiad fideo
- Astudiaeth Archwiliol i Effeithiau Delweddu ar Gywirdeb Canfod Twyll
- Canfod twyll amlfoddol mewn Fideos gyda Dysgu Dwfn
Bywgraffiad
- PhD mewn Cyfrifiadureg (Niwrowyddoniaeth Gyfrifiadurol) - Prifysgol Caint, Caergaint, DU
- MSc mewn Cyfrifiadureg Uwch (Cudd-wybodaeth Gyfrifiadurol): Rhagoriaeth - Prifysgol Caint, Caergaint, DU
- MSc Dylunio Cynnyrch a Systemau Peirianneg: Rhagoriaeth - Prifysgol yr Aegean, Syros, Gwlad Groeg
- Peirianneg Systemau Gwybodaeth a Chyfathrebu BEng - Prifysgol yr Aegean, Samos, Gwlad Groeg
Meysydd goruchwyliaeth
Ar hyn o bryd rwy'n edrych i gefnogi ceisiadau PhD ym maes canfod twyll awtomatig.
Mae gen i ddiddordeb hefyd mewn goruchwylio myfyrwyr PhD ym meysydd:
- Dadansoddiad fideo - micro-ymadroddion awtomatig, canfod ystumiau
- Dadansoddi testun - canfod twyll gan ddefnyddio dysgu peiriant / rhwydweithiau niwral
- Dadansoddiad llais - nodweddion ar gyfer canfod twyll
- Eyetrackers - echdynnu nodwedd ar gyfer dosbarthu
- HCI - sylw gweledol, amldasgio a datrys problemau - gan ddefnyddio tracio llygad a/neu niwroddelweddu
Goruchwyliaeth gyfredol
Asmail Muftah Muftah
Myfyriwr ymchwil