Reports de recercahttp://hdl.handle.net/2117/39252024-03-29T07:58:54Z2024-03-29T07:58:54ZIA generativa i ChatGPT. Impacte en l’aprenentatge universitariNadeu Camprubí, Climenthttp://hdl.handle.net/2117/4038232024-03-07T08:01:48Z2024-03-06T10:58:14ZIA generativa i ChatGPT. Impacte en l’aprenentatge universitari
Nadeu Camprubí, Climent
Es presenten de forma planera els elements bàsics de la intel·ligència artificial (IA) generativa. Sobretot es descriuen els models de llenguatge de gran capacitat, que són la base d’eines com el xatbot ChatGPT. A continuació es repassen les potencialitats i també les limitacions més rellevants dels sistemes basats en IA generativa, així com els riscs que comporta el seu ús. Finalment, es fa un intent de resposta parcial de
la pregunta sobre el seu impacte en les competències d’aprenentatge a la universitat.
Text escrit a principis de 2024 a partir de la ponència presentada per l’autor a la Taula Interuniversitària “Per a repensar l’experiència d’aprenentatge a la universitat a Catalunya”, celebrada a la UPC el dia 13 de juliol de 2023.
2024-03-06T10:58:14ZNadeu Camprubí, ClimentEs presenten de forma planera els elements bàsics de la intel·ligència artificial (IA) generativa. Sobretot es descriuen els models de llenguatge de gran capacitat, que són la base d’eines com el xatbot ChatGPT. A continuació es repassen les potencialitats i també les limitacions més rellevants dels sistemes basats en IA generativa, així com els riscs que comporta el seu ús. Finalment, es fa un intent de resposta parcial de
la pregunta sobre el seu impacte en les competències d’aprenentatge a la universitat.IST-033914-WINSOC: D4.4 System level network performance evaluationde Marinis, EnricoPagès Zamora, Alba MariaSilva Pereira, SilvanaYovanof, Gregory S.Pediaditaki, DimosthenisHasler, MartinAlireza, KhadiviLeonidas, Georgopouloshttp://hdl.handle.net/2117/3917952023-07-30T04:07:30Z2023-07-20T09:35:37ZIST-033914-WINSOC: D4.4 System level network performance evaluation
de Marinis, Enrico; Pagès Zamora, Alba Maria; Silva Pereira, Silvana; Yovanof, Gregory S.; Pediaditaki, Dimosthenis; Hasler, Martin; Alireza, Khadivi; Leonidas, Georgopoulos
This document constitutes the final outcome of the activities carried out in the WINSOC WP4, along with the interactions of this WP with WP3 and WP2. Although one objective of WP4 is to provide indications to the project implementation and validation activities (WP5 and WP6), the paradigms here presented as well as the results are self-standing and can be
viewed as a realistic evaluation point of the effectiveness of the system as a whole.
Confidencial
2023-07-20T09:35:37Zde Marinis, EnricoPagès Zamora, Alba MariaSilva Pereira, SilvanaYovanof, Gregory S.Pediaditaki, DimosthenisHasler, MartinAlireza, KhadiviLeonidas, GeorgopoulosThis document constitutes the final outcome of the activities carried out in the WINSOC WP4, along with the interactions of this WP with WP3 and WP2. Although one objective of WP4 is to provide indications to the project implementation and validation activities (WP5 and WP6), the paradigms here presented as well as the results are self-standing and can be
viewed as a realistic evaluation point of the effectiveness of the system as a whole.IST-2000-30148 I-METRA: D3.2 Implementation of relevant algorithmsRodríguez Fonollosa, JavierGaspa Maynou, RogerHeikkilä, Markku J.Kermoal, Jean PhilippeMajonen, KariMestre Pons, Francesc XavierPagès Zamora, Alba MariaPollard, AdamSchumacher, LaurentWiesel, AmiYlitalo, Juhahttp://hdl.handle.net/2117/3917702023-07-20T07:01:45Z2023-07-20T06:59:28ZIST-2000-30148 I-METRA: D3.2 Implementation of relevant algorithms
Rodríguez Fonollosa, Javier; Gaspa Maynou, Roger; Heikkilä, Markku J.; Kermoal, Jean Philippe; Majonen, Kari; Mestre Pons, Francesc Xavier; Pagès Zamora, Alba Maria; Pollard, Adam; Schumacher, Laurent; Wiesel, Ami; Ylitalo, Juha
This deliverable provides a high level description of the software developed within the I-METRA project following the selection reported in D3.1 "Design, Analysis and Selection of Suitable Algorithms".
2023-07-20T06:59:28ZRodríguez Fonollosa, JavierGaspa Maynou, RogerHeikkilä, Markku J.Kermoal, Jean PhilippeMajonen, KariMestre Pons, Francesc XavierPagès Zamora, Alba MariaPollard, AdamSchumacher, LaurentWiesel, AmiYlitalo, JuhaThis deliverable provides a high level description of the software developed within the I-METRA project following the selection reported in D3.1 "Design, Analysis and Selection of Suitable Algorithms".IST-2000-30148 I-METRA: D3.1 Design, analysis and selection of suitable algorithmsRodríguez Fonollosa, JavierGaspa Maynou, RogerHeikkilä, Markku J.Kermoal, Jean PhilippeMajonen, KariMestre Pons, Francesc XavierPagès Zamora, Alba MariaPollard, AdamSchumacher, LaurentWiesel, AmiYlitalo, Juhahttp://hdl.handle.net/2117/3917692023-07-20T06:45:07Z2023-07-20T06:42:44ZIST-2000-30148 I-METRA: D3.1 Design, analysis and selection of suitable algorithms
Rodríguez Fonollosa, Javier; Gaspa Maynou, Roger; Heikkilä, Markku J.; Kermoal, Jean Philippe; Majonen, Kari; Mestre Pons, Francesc Xavier; Pagès Zamora, Alba Maria; Pollard, Adam; Schumacher, Laurent; Wiesel, Ami; Ylitalo, Juha
This deliverable contains a description of the space-time coding algorithms to be simulated within the I-METRA project. Different families of algorithms have been selected and described in this document with the objective of evaluating their performance. One of the main objectives of the I-METRA project is to impact into the current standardisation efforts related to the introduction of Multiple Input Multiple Output (MIMO) configurations into the High Speed Downlink and Uplink Packet Access concepts of UMTS (HSDPA and HSUPA). This required a review of the current specifications for these systems and the analysis of the impact of the potential incorporation of the selected MIMO schemes.
2023-07-20T06:42:44ZRodríguez Fonollosa, JavierGaspa Maynou, RogerHeikkilä, Markku J.Kermoal, Jean PhilippeMajonen, KariMestre Pons, Francesc XavierPagès Zamora, Alba MariaPollard, AdamSchumacher, LaurentWiesel, AmiYlitalo, JuhaThis deliverable contains a description of the space-time coding algorithms to be simulated within the I-METRA project. Different families of algorithms have been selected and described in this document with the objective of evaluating their performance. One of the main objectives of the I-METRA project is to impact into the current standardisation efforts related to the introduction of Multiple Input Multiple Output (MIMO) configurations into the High Speed Downlink and Uplink Packet Access concepts of UMTS (HSDPA and HSUPA). This required a review of the current specifications for these systems and the analysis of the impact of the potential incorporation of the selected MIMO schemes.IST-2000-30148 I-METRA: D4 Performance evaluationHeikkilä, Markku J.Majonen, KariRodríguez Fonollosa, JavierPagès Zamora, Alba MariaYlitalo, JuhaTiirola, EsaPollard, Adamhttp://hdl.handle.net/2117/3917682023-07-20T06:37:21Z2023-07-20T06:29:31ZIST-2000-30148 I-METRA: D4 Performance evaluation
Heikkilä, Markku J.; Majonen, Kari; Rodríguez Fonollosa, Javier; Pagès Zamora, Alba Maria; Ylitalo, Juha; Tiirola, Esa; Pollard, Adam
This document considers the performance of multiantenna transmit/receive techniques in high-speed downlink and uplink packet access. The evaluation is done using both link and system level simulations by taking into account link adaptation and packet retransmissions. The document is based on the initial studies carried out in deliverables D3.1 and D3.2.
2023-07-20T06:29:31ZHeikkilä, Markku J.Majonen, KariRodríguez Fonollosa, JavierPagès Zamora, Alba MariaYlitalo, JuhaTiirola, EsaPollard, AdamThis document considers the performance of multiantenna transmit/receive techniques in high-speed downlink and uplink packet access. The evaluation is done using both link and system level simulations by taking into account link adaptation and packet retransmissions. The document is based on the initial studies carried out in deliverables D3.1 and D3.2.IST-2000-30148 I-METRA: D6.1 Implications in re-configurable systems beyond 3G (Part 1)Berger, Lars T.Rodríguez Fonollosa, JavierGarcía Ordóñez, LuísPagès Zamora, Alba MariaRamiro Moreno, Juanhttp://hdl.handle.net/2117/3909152023-07-14T10:59:18Z2023-07-14T10:57:41ZIST-2000-30148 I-METRA: D6.1 Implications in re-configurable systems beyond 3G (Part 1)
Berger, Lars T.; Rodríguez Fonollosa, Javier; García Ordóñez, Luís; Pagès Zamora, Alba Maria; Ramiro Moreno, Juan
In this activity MIMO HSDPA is evaluated as the UMTS evolution that could allow a combination of high bit rate services, coverage and mobility with a good trade-off between cost and performance. This evaluation requires the definition of an objective framework for comparison between competing air interface technologies for Systems beyond 3G, and should be carried out in cooperation with other IST projects. The deliverable is complemented by analytically assessing channel capacity in flat Rician- and Rayleigh fading when ideal proportional fast scheduling, optimal rate adaptation, and various transmit diversity techniques are used.
2023-07-14T10:57:41ZBerger, Lars T.Rodríguez Fonollosa, JavierGarcía Ordóñez, LuísPagès Zamora, Alba MariaRamiro Moreno, JuanIn this activity MIMO HSDPA is evaluated as the UMTS evolution that could allow a combination of high bit rate services, coverage and mobility with a good trade-off between cost and performance. This evaluation requires the definition of an objective framework for comparison between competing air interface technologies for Systems beyond 3G, and should be carried out in cooperation with other IST projects. The deliverable is complemented by analytically assessing channel capacity in flat Rician- and Rayleigh fading when ideal proportional fast scheduling, optimal rate adaptation, and various transmit diversity techniques are used.IST-2000-30148 I-METRA: D6.2 Implications in re-configurable systems beyond 3G (Part 2)Rodríguez Fonollosa, JavierPagès Zamora, Alba MariaTeyeb, Oumerhttp://hdl.handle.net/2117/3909002023-07-20T06:32:51Z2023-07-14T09:55:57ZIST-2000-30148 I-METRA: D6.2 Implications in re-configurable systems beyond 3G (Part 2)
Rodríguez Fonollosa, Javier; Pagès Zamora, Alba Maria; Teyeb, Oumer
This activity evaluates the extension of the bandwidth of the UTRA MIMO HSDPA concept to 20 MHz, which is precisely the bandwidth of HIPERLAN/2. This would allow a fair comparison between the performance of UTRA MIMO HSDPA and the enhanced HIPERLAN/2. The bandwidth expansion would be the consequence of multiplying the chip rate of the W-CDMA spreading by four, i.e., 3.84 x 4 = 15.36 Mcps. A higher bandwidth MIMO channel model is necessary and this will be developed based on the channel model already developed in WP2. High data rates are required to satisfy the ever-increasing application requirements in future wireless communication systems. Recent investigations have indicated that a peak data rate of up to 20Mbps per user in the DL may be required for satisfactory reception of bursty traffic. As the transmission powers (of both mobile terminals and base stations) are limited, higher data rates lead to the reduction of the effective coverage area of a cell. That is, only users that are close to the base station will be able to communicate with high data rates, while users far away from the base station will only be able to use low data rates.
2023-07-14T09:55:57ZRodríguez Fonollosa, JavierPagès Zamora, Alba MariaTeyeb, OumerThis activity evaluates the extension of the bandwidth of the UTRA MIMO HSDPA concept to 20 MHz, which is precisely the bandwidth of HIPERLAN/2. This would allow a fair comparison between the performance of UTRA MIMO HSDPA and the enhanced HIPERLAN/2. The bandwidth expansion would be the consequence of multiplying the chip rate of the W-CDMA spreading by four, i.e., 3.84 x 4 = 15.36 Mcps. A higher bandwidth MIMO channel model is necessary and this will be developed based on the channel model already developed in WP2. High data rates are required to satisfy the ever-increasing application requirements in future wireless communication systems. Recent investigations have indicated that a peak data rate of up to 20Mbps per user in the DL may be required for satisfactory reception of bursty traffic. As the transmission powers (of both mobile terminals and base stations) are limited, higher data rates lead to the reduction of the effective coverage area of a cell. That is, only users that are close to the base station will be able to communicate with high data rates, while users far away from the base station will only be able to use low data rates.Interpreting machine learning models for survival analysis: a study of cutaneous melanoma using the SEER databaseHernández Pérez, CarlosPachón García, CristianDelicado Useros, Pedro FranciscoVilaplana Besler, Verónicahttp://hdl.handle.net/2117/3900062023-06-30T11:30:15Z2023-06-30T11:28:18ZInterpreting machine learning models for survival analysis: a study of cutaneous melanoma using the SEER database
Hernández Pérez, Carlos; Pachón García, Cristian; Delicado Useros, Pedro Francisco; Vilaplana Besler, Verónica
In this study, we train and compare three types of machine learning algorithms for Survival Analysis: Random Survival Forest, DeepSurv and DeepHit, using the SEER database to model cutaneous malignant melanoma. Additionally, we employ SurvLIMEpy library, a Python package designed to provide explainability for survival machine learning models, to analyse feature importance. The results demonstrate that machine learning algorithms outperform the Cox Proportional Hazards Model. Our work underscores the importance of explainability methods for interpreting black-box models and provides insights into important features related to melanoma prognosis.
2023-06-30T11:28:18ZHernández Pérez, CarlosPachón García, CristianDelicado Useros, Pedro FranciscoVilaplana Besler, VerónicaIn this study, we train and compare three types of machine learning algorithms for Survival Analysis: Random Survival Forest, DeepSurv and DeepHit, using the SEER database to model cutaneous malignant melanoma. Additionally, we employ SurvLIMEpy library, a Python package designed to provide explainability for survival machine learning models, to analyse feature importance. The results demonstrate that machine learning algorithms outperform the Cox Proportional Hazards Model. Our work underscores the importance of explainability methods for interpreting black-box models and provides insights into important features related to melanoma prognosis.The 2021 volcanic eruption in La Palma Island and its impact on ionospheric scintillation as measured from GNSS reference stations, GNSS-R, and GNSS-ROMolina Ordóñez, CarlosBoudriki Semlali, Badr EddineHyuk, ParkCamps Carmona, Adriano Joséhttp://hdl.handle.net/2117/3899232023-10-29T10:22:11Z2023-06-29T09:36:12ZThe 2021 volcanic eruption in La Palma Island and its impact on ionospheric scintillation as measured from GNSS reference stations, GNSS-R, and GNSS-RO
Molina Ordóñez, Carlos; Boudriki Semlali, Badr Eddine; Hyuk, Park; Camps Carmona, Adriano José
Ionospheric disturbances induced by seismic activity have been studied in the last years by many authors, showing an impact both before and after the occurrence of earthquakes. In this study, the ionospheric scintillation produced by the 2021 La Palma volcano eruption is analyzed. The "Cumbre Vieja" volcano was active from September 19th to December 13th, 2021, and many magnitude 3–4 earthquakes were recorded, with some of them reaching magnitude 5. In this study the three methods: GNSS reference monitoring, GNSS Reflectometry (GNSS-R) from NASA CYGNSS, and GNSS Radio Occultation (GNSS-RO) from COSMIC and Spire constellations, are used, allowing us to compare and evaluate their performance in the same conditions. To compare the seismic activity with ionospheric scintillation, earthquakes’ generated energy, and percentile 95 % of the intensity scintillation parameter (S4), measurements have been computed every 6 h intervals for the whole duration of the volcanic eruption. GNSS-RO has shown the best correlation between earthquakes’ energy and S4, with values up to 0.09 when the perturbations occur around 18 h after the seismic activity. GNSS reference monitoring stations data also shows some correlation 18 h after and 7–8 days after. As expected, GNSS-R is the one that shows the smallest correlation, as the ionospheric signatures get masked by the signature of the surface where the reflection is taking place. Additionally, as expected as well, the three methods show a smaller correlation during the week before earthquakes.
2023-06-29T09:36:12ZMolina Ordóñez, CarlosBoudriki Semlali, Badr EddineHyuk, ParkCamps Carmona, Adriano JoséIonospheric disturbances induced by seismic activity have been studied in the last years by many authors, showing an impact both before and after the occurrence of earthquakes. In this study, the ionospheric scintillation produced by the 2021 La Palma volcano eruption is analyzed. The "Cumbre Vieja" volcano was active from September 19th to December 13th, 2021, and many magnitude 3–4 earthquakes were recorded, with some of them reaching magnitude 5. In this study the three methods: GNSS reference monitoring, GNSS Reflectometry (GNSS-R) from NASA CYGNSS, and GNSS Radio Occultation (GNSS-RO) from COSMIC and Spire constellations, are used, allowing us to compare and evaluate their performance in the same conditions. To compare the seismic activity with ionospheric scintillation, earthquakes’ generated energy, and percentile 95 % of the intensity scintillation parameter (S4), measurements have been computed every 6 h intervals for the whole duration of the volcanic eruption. GNSS-RO has shown the best correlation between earthquakes’ energy and S4, with values up to 0.09 when the perturbations occur around 18 h after the seismic activity. GNSS reference monitoring stations data also shows some correlation 18 h after and 7–8 days after. As expected, GNSS-R is the one that shows the smallest correlation, as the ionospheric signatures get masked by the signature of the surface where the reflection is taking place. Additionally, as expected as well, the three methods show a smaller correlation during the week before earthquakes.Portable kit for detecting trapped and buried people in ruins and avalanches, RESCUECELL: D2.3 Report and source code of final MT locationSilva Pereira, SilvanaGiménez Febrer, Pedro JuanFontdevila Olivé, RogerAtzeni, ItaloPagès Zamora, Alba MariaVidal Manzano, Joséhttp://hdl.handle.net/2117/3877822023-05-24T11:31:05Z2023-05-24T11:29:31ZPortable kit for detecting trapped and buried people in ruins and avalanches, RESCUECELL: D2.3 Report and source code of final MT location
Silva Pereira, Silvana; Giménez Febrer, Pedro Juan; Fontdevila Olivé, Roger; Atzeni, Italo; Pagès Zamora, Alba Maria; Vidal Manzano, José
RESCUECELL. Seventh Framework Programme Research for the Benefit of the SMEs. Grant Agreement Number 315007
Deliverable: D2.3 Report and source code of final MT location
2023-05-24T11:29:31ZSilva Pereira, SilvanaGiménez Febrer, Pedro JuanFontdevila Olivé, RogerAtzeni, ItaloPagès Zamora, Alba MariaVidal Manzano, JoséRESCUECELL. Seventh Framework Programme Research for the Benefit of the SMEs. Grant Agreement Number 315007
Deliverable: D2.3 Report and source code of final MT location