ݺߣshows by User: balazshidasi / http://www.slideshare.net/images/logo.gif ݺߣshows by User: balazshidasi / Wed, 15 Nov 2023 13:54:05 GMT ݺߣShare feed for ݺߣshows by User: balazshidasi Egyedi termék kreatívok tömeges gyártása generatív AI segítségével /slideshow/egyedi-termk-kreatvok-tmeges-gyrtsa-generatv-ai-segtsgvel/263459147 genaimeetup20231114-231115135405-b9a22d53
UPDATE: Typo on the 8th slide, last line should be (slides can't be modified on slideshare): grad(log(p_gamma(x|y))) = (1-gamma)*grad(log(p(x))) + gamma*grad(log(p(x|y))) My presentation on using generative AI for creative generation for e-commerce. Presented on 14 November 2023 at the TECH meetup series organized by Gravity R&D, a Taboola company. ݺߣs are in Hungarian. Title/abstract in English: Mass production of unique product creatives with generative AI ----- The probability of a user clicking on an online advertisement is greatly influenced by creative's look. Traditional brand level campaigns require only a few creatives that can be produced by humans. However product level recommendations require creatives for every single product. Producing these using human work is infeasible at scale, thus they are often shown in front of simple (e.g. white) backgrounds. This presentation showcases a solution based on generative AI that allows placing products in different environments, which makes the creatives more appealing. I'll talk about the challenges of this approach along with potential solutions, as well as the initial results of our live test. Eredeti absztrakt: Az online hirdetések megjelenése nagyban befolyásolja a rákattintás valószínűségét. A tradicionális márka szinten targetált kampányokhoz szükséges egy-két kreatív/banner legyártása még emberi erőforrás igénybevételével is megoldható. Termék szintű ajánlás esetén viszont minden egyes termékhez külön kreatívra van szükség, akár több felbontásban. Nagyszámú kreatív legyártása emberi erővel lassú és drága, ezért gyakori megközelítés a terméket valamilyen egyszerű, például egyszínű, háttér előtt megjeleníteni. Az előadás során bemutatunk egy generatív AI technológián alapuló megoldást, ami lehetővé teszi, hogy a termékeket különféle környezetekben jelenítsük meg, és így érdekesebbé/vonzóbbá tegyük a kreatívokat. Szót ejtünk a megközelítés nehézségeiről, lehetséges megoldásokról, és a módszer hatékonyságát vizsgáló mérésünk előzetes eredményeiről.]]>
UPDATE: Typo on the 8th slide, last line should be (slides can't be modified on slideshare): grad(log(p_gamma(x|y))) = (1-gamma)*grad(log(p(x))) + gamma*grad(log(p(x|y))) My presentation on using generative AI for creative generation for e-commerce. Presented on 14 November 2023 at the TECH meetup series organized by Gravity R&D, a Taboola company. ݺߣs are in Hungarian. Title/abstract in English: Mass production of unique product creatives with generative AI ----- The probability of a user clicking on an online advertisement is greatly influenced by creative's look. Traditional brand level campaigns require only a few creatives that can be produced by humans. However product level recommendations require creatives for every single product. Producing these using human work is infeasible at scale, thus they are often shown in front of simple (e.g. white) backgrounds. This presentation showcases a solution based on generative AI that allows placing products in different environments, which makes the creatives more appealing. I'll talk about the challenges of this approach along with potential solutions, as well as the initial results of our live test. Eredeti absztrakt: Az online hirdetések megjelenése nagyban befolyásolja a rákattintás valószínűségét. A tradicionális márka szinten targetált kampányokhoz szükséges egy-két kreatív/banner legyártása még emberi erőforrás igénybevételével is megoldható. Termék szintű ajánlás esetén viszont minden egyes termékhez külön kreatívra van szükség, akár több felbontásban. Nagyszámú kreatív legyártása emberi erővel lassú és drága, ezért gyakori megközelítés a terméket valamilyen egyszerű, például egyszínű, háttér előtt megjeleníteni. Az előadás során bemutatunk egy generatív AI technológián alapuló megoldást, ami lehetővé teszi, hogy a termékeket különféle környezetekben jelenítsük meg, és így érdekesebbé/vonzóbbá tegyük a kreatívokat. Szót ejtünk a megközelítés nehézségeiről, lehetséges megoldásokról, és a módszer hatékonyságát vizsgáló mérésünk előzetes eredményeiről.]]> Wed, 15 Nov 2023 13:54:05 GMT /slideshow/egyedi-termk-kreatvok-tmeges-gyrtsa-generatv-ai-segtsgvel/263459147 balazshidasi@slideshare.net(balazshidasi) Egyedi termék kreatívok tömeges gyártása generatív AI segítségével balazshidasi UPDATE: Typo on the 8th slide, last line should be (slides can't be modified on slideshare): grad(log(p_gamma(x|y))) = (1-gamma)*grad(log(p(x))) + gamma*grad(log(p(x|y))) My presentation on using generative AI for creative generation for e-commerce. Presented on 14 November 2023 at the TECH meetup series organized by Gravity R&D, a Taboola company. ݺߣs are in Hungarian. Title/abstract in English: Mass production of unique product creatives with generative AI ----- The probability of a user clicking on an online advertisement is greatly influenced by creative's look. Traditional brand level campaigns require only a few creatives that can be produced by humans. However product level recommendations require creatives for every single product. Producing these using human work is infeasible at scale, thus they are often shown in front of simple (e.g. white) backgrounds. This presentation showcases a solution based on generative AI that allows placing products in different environments, which makes the creatives more appealing. I'll talk about the challenges of this approach along with potential solutions, as well as the initial results of our live test. Eredeti absztrakt: Az online hirdetések megjelenése nagyban befolyásolja a rákattintás valószínűségét. A tradicionális márka szinten targetált kampányokhoz szükséges egy-két kreatív/banner legyártása még emberi erőforrás igénybevételével is megoldható. Termék szintű ajánlás esetén viszont minden egyes termékhez külön kreatívra van szükség, akár több felbontásban. Nagyszámú kreatív legyártása emberi erővel lassú és drága, ezért gyakori megközelítés a terméket valamilyen egyszerű, például egyszínű, háttér előtt megjeleníteni. Az előadás során bemutatunk egy generatív AI technológián alapuló megoldást, ami lehetővé teszi, hogy a termékeket különféle környezetekben jelenítsük meg, és így érdekesebbé/vonzóbbá tegyük a kreatívokat. Szót ejtünk a megközelítés nehézségeiről, lehetséges megoldásokról, és a módszer hatékonyságát vizsgáló mérésünk előzetes eredményeiről. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/genaimeetup20231114-231115135405-b9a22d53-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> UPDATE: Typo on the 8th slide, last line should be (slides can&#39;t be modified on slideshare): grad(log(p_gamma(x|y))) = (1-gamma)*grad(log(p(x))) + gamma*grad(log(p(x|y))) My presentation on using generative AI for creative generation for e-commerce. Presented on 14 November 2023 at the TECH meetup series organized by Gravity R&amp;D, a Taboola company. ݺߣs are in Hungarian. Title/abstract in English: Mass production of unique product creatives with generative AI ----- The probability of a user clicking on an online advertisement is greatly influenced by creative&#39;s look. Traditional brand level campaigns require only a few creatives that can be produced by humans. However product level recommendations require creatives for every single product. Producing these using human work is infeasible at scale, thus they are often shown in front of simple (e.g. white) backgrounds. This presentation showcases a solution based on generative AI that allows placing products in different environments, which makes the creatives more appealing. I&#39;ll talk about the challenges of this approach along with potential solutions, as well as the initial results of our live test. Eredeti absztrakt: Az online hirdetések megjelenése nagyban befolyásolja a rákattintás valószínűségét. A tradicionális márka szinten targetált kampányokhoz szükséges egy-két kreatív/banner legyártása még emberi erőforrás igénybevételével is megoldható. Termék szintű ajánlás esetén viszont minden egyes termékhez külön kreatívra van szükség, akár több felbontásban. Nagyszámú kreatív legyártása emberi erővel lassú és drága, ezért gyakori megközelítés a terméket valamilyen egyszerű, például egyszínű, háttér előtt megjeleníteni. Az előadás során bemutatunk egy generatív AI technológián alapuló megoldást, ami lehetővé teszi, hogy a termékeket különféle környezetekben jelenítsük meg, és így érdekesebbé/vonzóbbá tegyük a kreatívokat. Szót ejtünk a megközelítés nehézségeiről, lehetséges megoldásokról, és a módszer hatékonyságát vizsgáló mérésünk előzetes eredményeiről.
Egyedi term辿k kreat鱈vok t旦meges gy叩rt叩sa generat鱈v AI seg鱈ts辿g辿vel from Bal叩zs Hidasi
]]> 59 0 https://cdn.slidesharecdn.com/ss_thumbnails/genaimeetup20231114-231115135405-b9a22d53-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 The Effect of Third Party Implementations on Reproducibility /slideshow/the-effect-of-third-party-implementations-on-reproducibility/261174582 thirdpartyevalsplitslides-230920031211-dbcfe40c
Presentation of "The Effect of Third Party Implementations on Reproducibility" paper from RecSys 2023. Abstract: Reproducibility of recommender systems research has come under scrutiny during recent years. Along with works focusing on repeating experiments with certain algorithms, the research community has also started discussing various aspects of evaluation and how these affect reproducibility. We add a novel angle to this discussion by examining how unofficial third-party implementations could benefit or hinder reproducibility. Besides giving a general overview, we thoroughly examine six third-party implementations of a popular recommender algorithm and compare them to the official version on five public datasets. In the light of our alarming findings we aim to draw the attention of the research community to this neglected aspect of reproducibility.]]>
Presentation of "The Effect of Third Party Implementations on Reproducibility" paper from RecSys 2023. Abstract: Reproducibility of recommender systems research has come under scrutiny during recent years. Along with works focusing on repeating experiments with certain algorithms, the research community has also started discussing various aspects of evaluation and how these affect reproducibility. We add a novel angle to this discussion by examining how unofficial third-party implementations could benefit or hinder reproducibility. Besides giving a general overview, we thoroughly examine six third-party implementations of a popular recommender algorithm and compare them to the official version on five public datasets. In the light of our alarming findings we aim to draw the attention of the research community to this neglected aspect of reproducibility.]]> Wed, 20 Sep 2023 03:12:10 GMT /slideshow/the-effect-of-third-party-implementations-on-reproducibility/261174582 balazshidasi@slideshare.net(balazshidasi) The Effect of Third Party Implementations on Reproducibility balazshidasi Presentation of "The Effect of Third Party Implementations on Reproducibility" paper from RecSys 2023. Abstract: Reproducibility of recommender systems research has come under scrutiny during recent years. Along with works focusing on repeating experiments with certain algorithms, the research community has also started discussing various aspects of evaluation and how these affect reproducibility. We add a novel angle to this discussion by examining how unofficial third-party implementations could benefit or hinder reproducibility. Besides giving a general overview, we thoroughly examine six third-party implementations of a popular recommender algorithm and compare them to the official version on five public datasets. In the light of our alarming findings we aim to draw the attention of the research community to this neglected aspect of reproducibility. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/thirdpartyevalsplitslides-230920031211-dbcfe40c-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Presentation of &quot;The Effect of Third Party Implementations on Reproducibility&quot; paper from RecSys 2023. Abstract: Reproducibility of recommender systems research has come under scrutiny during recent years. Along with works focusing on repeating experiments with certain algorithms, the research community has also started discussing various aspects of evaluation and how these affect reproducibility. We add a novel angle to this discussion by examining how unofficial third-party implementations could benefit or hinder reproducibility. Besides giving a general overview, we thoroughly examine six third-party implementations of a popular recommender algorithm and compare them to the official version on five public datasets. In the light of our alarming findings we aim to draw the attention of the research community to this neglected aspect of reproducibility.
The Effect of Third Party Implementations on Reproducibility from Bal叩zs Hidasi
]]> 145 0 https://cdn.slidesharecdn.com/ss_thumbnails/thirdpartyevalsplitslides-230920031211-dbcfe40c-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 GRU4Rec v2 - Recurrent Neural Networks with Top-k Gains for Session-based Recommendations /slideshow/gru4rec-v2-recurrent-neural-networks-with-topk-gains-for-sessionbased-recommendations/120561602 gru4recv2cikm18-181024131645
ݺߣs of my presentation at CIKM2018 about version 2 of the GRU4Rec algorithm, a recurrent neural network based algorithm for the session-based recommendation task. We discuss sampling strategies and introduce additional sampling to the algorithm. We also redesign the loss function to cope with additional sampling. The resulting BPR-max loss function is able to efficiently handle many negative samples without encountering the vanishing gradient problem. We also introduce constrained embeddings which speeds up the conversion of item representations and reduces memory usage by a factor of 4. These improvements increase offline measures up to 52%. In the talk we also discuss online A/B test and the implications of long time observations. Most of these observations are exclusive to this talk and are not in the paper. You can access the preprint version of the paper on arXiv: https://arxiv.org/abs/1706.03847 The code is available on GitHub: https://github.com/hidasib/GRU4Rec ]]>
ݺߣs of my presentation at CIKM2018 about version 2 of the GRU4Rec algorithm, a recurrent neural network based algorithm for the session-based recommendation task. We discuss sampling strategies and introduce additional sampling to the algorithm. We also redesign the loss function to cope with additional sampling. The resulting BPR-max loss function is able to efficiently handle many negative samples without encountering the vanishing gradient problem. We also introduce constrained embeddings which speeds up the conversion of item representations and reduces memory usage by a factor of 4. These improvements increase offline measures up to 52%. In the talk we also discuss online A/B test and the implications of long time observations. Most of these observations are exclusive to this talk and are not in the paper. You can access the preprint version of the paper on arXiv: https://arxiv.org/abs/1706.03847 The code is available on GitHub: https://github.com/hidasib/GRU4Rec ]]> Wed, 24 Oct 2018 13:16:45 GMT /slideshow/gru4rec-v2-recurrent-neural-networks-with-topk-gains-for-sessionbased-recommendations/120561602 balazshidasi@slideshare.net(balazshidasi) GRU4Rec v2 - Recurrent Neural Networks with Top-k Gains for Session-based Recommendations balazshidasi ݺߣs of my presentation at CIKM2018 about version 2 of the GRU4Rec algorithm, a recurrent neural network based algorithm for the session-based recommendation task. We discuss sampling strategies and introduce additional sampling to the algorithm. We also redesign the loss function to cope with additional sampling. The resulting BPR-max loss function is able to efficiently handle many negative samples without encountering the vanishing gradient problem. We also introduce constrained embeddings which speeds up the conversion of item representations and reduces memory usage by a factor of 4. These improvements increase offline measures up to 52%. In the talk we also discuss online A/B test and the implications of long time observations. Most of these observations are exclusive to this talk and are not in the paper. You can access the preprint version of the paper on arXiv: https://arxiv.org/abs/1706.03847 The code is available on GitHub: https://github.com/hidasib/GRU4Rec <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/gru4recv2cikm18-181024131645-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> ݺߣs of my presentation at CIKM2018 about version 2 of the GRU4Rec algorithm, a recurrent neural network based algorithm for the session-based recommendation task. We discuss sampling strategies and introduce additional sampling to the algorithm. We also redesign the loss function to cope with additional sampling. The resulting BPR-max loss function is able to efficiently handle many negative samples without encountering the vanishing gradient problem. We also introduce constrained embeddings which speeds up the conversion of item representations and reduces memory usage by a factor of 4. These improvements increase offline measures up to 52%. In the talk we also discuss online A/B test and the implications of long time observations. Most of these observations are exclusive to this talk and are not in the paper. You can access the preprint version of the paper on arXiv: https://arxiv.org/abs/1706.03847 The code is available on GitHub: https://github.com/hidasib/GRU4Rec
GRU4Rec v2 - Recurrent Neural Networks with Top-k Gains for Session-based Recommendations from Bal叩zs Hidasi
]]> 3455 2 https://cdn.slidesharecdn.com/ss_thumbnails/gru4recv2cikm18-181024131645-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Deep Learning in Recommender Systems - RecSys Summer School 2017 /slideshow/deep-learning-in-recommender-systems-recsys-summer-school-2017/79047428 dltutorialrecsys17deeplearning-170822092943
This is the presentation accompanying my tutorial about deep learning methods in the recommender systems domain. The tutorial consists of a brief general overview of deep learning and the introduction of the four most prominent research direction of DL in recsys as of 2017. Presented during RecSys Summer School 2017 in Bolzano, Italy. ]]>
This is the presentation accompanying my tutorial about deep learning methods in the recommender systems domain. The tutorial consists of a brief general overview of deep learning and the introduction of the four most prominent research direction of DL in recsys as of 2017. Presented during RecSys Summer School 2017 in Bolzano, Italy. ]]> Tue, 22 Aug 2017 09:29:43 GMT /slideshow/deep-learning-in-recommender-systems-recsys-summer-school-2017/79047428 balazshidasi@slideshare.net(balazshidasi) Deep Learning in Recommender Systems - RecSys Summer School 2017 balazshidasi This is the presentation accompanying my tutorial about deep learning methods in the recommender systems domain. The tutorial consists of a brief general overview of deep learning and the introduction of the four most prominent research direction of DL in recsys as of 2017. Presented during RecSys Summer School 2017 in Bolzano, Italy. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/dltutorialrecsys17deeplearning-170822092943-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> This is the presentation accompanying my tutorial about deep learning methods in the recommender systems domain. The tutorial consists of a brief general overview of deep learning and the introduction of the four most prominent research direction of DL in recsys as of 2017. Presented during RecSys Summer School 2017 in Bolzano, Italy.
Deep Learning in Recommender Systems - RecSys Summer School 2017 from Bal叩zs Hidasi
]]> 11457 30 https://cdn.slidesharecdn.com/ss_thumbnails/dltutorialrecsys17deeplearning-170822092943-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Parallel Recurrent Neural Network Architectures for Feature-rich Session-based Recommendations /slideshow/parallel-recurrent-neural-network-architectures-for-featurerich-sessionbased-recommendations/66144702 prnnslides-160918182933
ݺߣs for my RecSys 2016 talk on integrating image and textual information into session based recommendations using novel parallel RNN architectures. Link to the paper: http://www.hidasi.eu/en/publications.html#p_rnn_recsys16]]>
ݺߣs for my RecSys 2016 talk on integrating image and textual information into session based recommendations using novel parallel RNN architectures. Link to the paper: http://www.hidasi.eu/en/publications.html#p_rnn_recsys16]]> Sun, 18 Sep 2016 18:29:33 GMT /slideshow/parallel-recurrent-neural-network-architectures-for-featurerich-sessionbased-recommendations/66144702 balazshidasi@slideshare.net(balazshidasi) Parallel Recurrent Neural Network Architectures for Feature-rich Session-based Recommendations balazshidasi ݺߣs for my RecSys 2016 talk on integrating image and textual information into session based recommendations using novel parallel RNN architectures. Link to the paper: http://www.hidasi.eu/en/publications.html#p_rnn_recsys16 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/prnnslides-160918182933-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> ݺߣs for my RecSys 2016 talk on integrating image and textual information into session based recommendations using novel parallel RNN architectures. Link to the paper: http://www.hidasi.eu/en/publications.html#p_rnn_recsys16
Parallel Recurrent Neural Network Architectures for Feature-rich Session-based Recommendations from Bal叩zs Hidasi
]]> 2278 3 https://cdn.slidesharecdn.com/ss_thumbnails/prnnslides-160918182933-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Context aware factorization methods for implicit feedback based recommendation problems (HUN) /slideshow/context-aware-factorization-methods-for-implicit-feedback-based-recommendation-problems-hun/63522443 context-awarefactorizationmethodsforimplicitfeedbackbasedrecommendationproblems-160628120212
ݺߣs I prepared for defending my PhD dissertation on context-aware factorization methods for implicit-feedback based recommendations. Dissertation (in English) can be accessed here: http://hidasi.eu/content/phd.pdf ݺߣs are in Hungarian.]]>
ݺߣs I prepared for defending my PhD dissertation on context-aware factorization methods for implicit-feedback based recommendations. Dissertation (in English) can be accessed here: http://hidasi.eu/content/phd.pdf ݺߣs are in Hungarian.]]> Tue, 28 Jun 2016 12:02:12 GMT /slideshow/context-aware-factorization-methods-for-implicit-feedback-based-recommendation-problems-hun/63522443 balazshidasi@slideshare.net(balazshidasi) Context aware factorization methods for implicit feedback based recommendation problems (HUN) balazshidasi ݺߣs I prepared for defending my PhD dissertation on context-aware factorization methods for implicit-feedback based recommendations. Dissertation (in English) can be accessed here: http://hidasi.eu/content/phd.pdf ݺߣs are in Hungarian. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/context-awarefactorizationmethodsforimplicitfeedbackbasedrecommendationproblems-160628120212-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> ݺߣs I prepared for defending my PhD dissertation on context-aware factorization methods for implicit-feedback based recommendations. Dissertation (in English) can be accessed here: http://hidasi.eu/content/phd.pdf ݺߣs are in Hungarian.
Context aware factorization methods for implicit feedback based recommendation problems (HUN) from Bal叩zs Hidasi
]]> 527 10 https://cdn.slidesharecdn.com/ss_thumbnails/context-awarefactorizationmethodsforimplicitfeedbackbasedrecommendationproblems-160628120212-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Deep learning to the rescue - �solving long standing problems of recommender systems /slideshow/deep-learning-to-the-rescue-solving-long-standing-problems-of-recommender-systems/61968320 dltalkrecsysmeetup20160512v2-160512234222
I gave this talk at the 1st Budapest RecSys and Personalization Meetup about using deep learning to solve long standing problems of recommender systems. I also presented our approach on using RNNs for session-based recommendations in details.]]>
I gave this talk at the 1st Budapest RecSys and Personalization Meetup about using deep learning to solve long standing problems of recommender systems. I also presented our approach on using RNNs for session-based recommendations in details.]]> Thu, 12 May 2016 23:42:22 GMT /slideshow/deep-learning-to-the-rescue-solving-long-standing-problems-of-recommender-systems/61968320 balazshidasi@slideshare.net(balazshidasi) Deep learning to the rescue - �solving long standing problems of recommender systems balazshidasi I gave this talk at the 1st Budapest RecSys and Personalization Meetup about using deep learning to solve long standing problems of recommender systems. I also presented our approach on using RNNs for session-based recommendations in details. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/dltalkrecsysmeetup20160512v2-160512234222-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> I gave this talk at the 1st Budapest RecSys and Personalization Meetup about using deep learning to solve long standing problems of recommender systems. I also presented our approach on using RNNs for session-based recommendations in details.
Deep learning to the rescue - solving long standing problems of recommender systems from Bal叩zs Hidasi
]]> 13557 8 https://cdn.slidesharecdn.com/ss_thumbnails/dltalkrecsysmeetup20160512v2-160512234222-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Deep learning: the future of recommendations /slideshow/deep-learning-the-future-of-recommendations/61241457 deeplearningtalkstartupsafary20160421-160422150918
An informative talk about deep learning and its potential uses in recommender systems. Presented at the Budapest Startup Safary, 21 April, 2016. The breakthroughs of the last decade in neural network research and the quick increasing of computational power resulted in the revival of deep neural networks and the field focusing on their training: deep learning. Deep learning methods have succeeded in complex tasks where other machine learning methods have failed, such as computer vision and natural language processing. Recently deep learning has began to gain ground in recommender systems as well. This talk introduces deep learning and its applications, with emphasis on how deep learning methods can solve long standing recommendation problems.]]>
An informative talk about deep learning and its potential uses in recommender systems. Presented at the Budapest Startup Safary, 21 April, 2016. The breakthroughs of the last decade in neural network research and the quick increasing of computational power resulted in the revival of deep neural networks and the field focusing on their training: deep learning. Deep learning methods have succeeded in complex tasks where other machine learning methods have failed, such as computer vision and natural language processing. Recently deep learning has began to gain ground in recommender systems as well. This talk introduces deep learning and its applications, with emphasis on how deep learning methods can solve long standing recommendation problems.]]> Fri, 22 Apr 2016 15:09:18 GMT /slideshow/deep-learning-the-future-of-recommendations/61241457 balazshidasi@slideshare.net(balazshidasi) Deep learning: the future of recommendations balazshidasi An informative talk about deep learning and its potential uses in recommender systems. Presented at the Budapest Startup Safary, 21 April, 2016. The breakthroughs of the last decade in neural network research and the quick increasing of computational power resulted in the revival of deep neural networks and the field focusing on their training: deep learning. Deep learning methods have succeeded in complex tasks where other machine learning methods have failed, such as computer vision and natural language processing. Recently deep learning has began to gain ground in recommender systems as well. This talk introduces deep learning and its applications, with emphasis on how deep learning methods can solve long standing recommendation problems. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/deeplearningtalkstartupsafary20160421-160422150918-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> An informative talk about deep learning and its potential uses in recommender systems. Presented at the Budapest Startup Safary, 21 April, 2016. The breakthroughs of the last decade in neural network research and the quick increasing of computational power resulted in the revival of deep neural networks and the field focusing on their training: deep learning. Deep learning methods have succeeded in complex tasks where other machine learning methods have failed, such as computer vision and natural language processing. Recently deep learning has began to gain ground in recommender systems as well. This talk introduces deep learning and its applications, with emphasis on how deep learning methods can solve long standing recommendation problems.
Deep learning: the future of recommendations from Bal叩zs Hidasi
]]> 15140 12 https://cdn.slidesharecdn.com/ss_thumbnails/deeplearningtalkstartupsafary20160421-160422150918-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Context-aware preference modeling with factorization /slideshow/contextaware-preference-modeling-with-factorization/52981720 dsrecsys15slides-150920120152-lva1-app6892
This talk was presented at the Doctoral Symposium of RecSys'15. It is a summary of the core part of my PhD research in the last few years. The research revolves around solving the implicit feedback based context-aware recommendation problem with factorization. Associated paper: http://dl.acm.org/citation.cfm?id=2796543 Details of presented algorithms/methods (public versions available on http://hidasi.eu): iTALS: http://link.springer.com/chapter/10.1007/978-3-642-33486-3_5 iTALSx: http://www.infocommunications.hu/documents/169298/1025723/InfocomJ_2014_4_5_Hidasi.pdf ALS-CG/CD: http://link.springer.com/article/10.1007/s10115-015-0863-2 GFF: http://link.springer.com/article/10.1007/s10618-015-0417-y]]>
This talk was presented at the Doctoral Symposium of RecSys'15. It is a summary of the core part of my PhD research in the last few years. The research revolves around solving the implicit feedback based context-aware recommendation problem with factorization. Associated paper: http://dl.acm.org/citation.cfm?id=2796543 Details of presented algorithms/methods (public versions available on http://hidasi.eu): iTALS: http://link.springer.com/chapter/10.1007/978-3-642-33486-3_5 iTALSx: http://www.infocommunications.hu/documents/169298/1025723/InfocomJ_2014_4_5_Hidasi.pdf ALS-CG/CD: http://link.springer.com/article/10.1007/s10115-015-0863-2 GFF: http://link.springer.com/article/10.1007/s10618-015-0417-y]]> Sun, 20 Sep 2015 12:01:52 GMT /slideshow/contextaware-preference-modeling-with-factorization/52981720 balazshidasi@slideshare.net(balazshidasi) Context-aware preference modeling with factorization balazshidasi This talk was presented at the Doctoral Symposium of RecSys'15. It is a summary of the core part of my PhD research in the last few years. The research revolves around solving the implicit feedback based context-aware recommendation problem with factorization. Associated paper: http://dl.acm.org/citation.cfm?id=2796543 Details of presented algorithms/methods (public versions available on http://hidasi.eu): iTALS: http://link.springer.com/chapter/10.1007/978-3-642-33486-3_5 iTALSx: http://www.infocommunications.hu/documents/169298/1025723/InfocomJ_2014_4_5_Hidasi.pdf ALS-CG/CD: http://link.springer.com/article/10.1007/s10115-015-0863-2 GFF: http://link.springer.com/article/10.1007/s10618-015-0417-y <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/dsrecsys15slides-150920120152-lva1-app6892-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> This talk was presented at the Doctoral Symposium of RecSys&#39;15. It is a summary of the core part of my PhD research in the last few years. The research revolves around solving the implicit feedback based context-aware recommendation problem with factorization. Associated paper: http://dl.acm.org/citation.cfm?id=2796543 Details of presented algorithms/methods (public versions available on http://hidasi.eu): iTALS: http://link.springer.com/chapter/10.1007/978-3-642-33486-3_5 iTALSx: http://www.infocommunications.hu/documents/169298/1025723/InfocomJ_2014_4_5_Hidasi.pdf ALS-CG/CD: http://link.springer.com/article/10.1007/s10115-015-0863-2 GFF: http://link.springer.com/article/10.1007/s10618-015-0417-y
Context-aware preference modeling with factorization from Bal叩zs Hidasi
]]> 1664 7 https://cdn.slidesharecdn.com/ss_thumbnails/dsrecsys15slides-150920120152-lva1-app6892-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Approximate modeling of continuous context in factorization algorithms (CaRR14 presentation) /slideshow/fuzzy-context-carr14/33463797 fuzzycontextcarr14-140413084344-phpapp02
]]>
]]> Sun, 13 Apr 2014 08:43:43 GMT /slideshow/fuzzy-context-carr14/33463797 balazshidasi@slideshare.net(balazshidasi) Approximate modeling of continuous context in factorization algorithms (CaRR14 presentation) balazshidasi <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/fuzzycontextcarr14-140413084344-phpapp02-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br>
Approximate modeling of continuous context in factorization algorithms (CaRR14 presentation) from Bal叩zs Hidasi
]]> 1365 4 https://cdn.slidesharecdn.com/ss_thumbnails/fuzzycontextcarr14-140413084344-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Utilizing additional information in factorization methods (research overview, April 2014) /slideshow/utilizing-additional-information-in-factorization-methods-research-overview-april-2014/33450437 tudprezi-140412145045-phpapp01
This presentation contains the main points of my recommender systems related research. It describes the arc of my research starting from improving matrix factorization, through the developement of my context-aware algorithms & addressing scalability issues to developing a general factorization framework & dealing with context dimension modeling. The slides were presented at the Delft University of Technology where I was invited to give this introductory talk as part of the collaboration between participiants of the CrowdRec project. The presentation was given on 11th April 2014.]]>
This presentation contains the main points of my recommender systems related research. It describes the arc of my research starting from improving matrix factorization, through the developement of my context-aware algorithms & addressing scalability issues to developing a general factorization framework & dealing with context dimension modeling. The slides were presented at the Delft University of Technology where I was invited to give this introductory talk as part of the collaboration between participiants of the CrowdRec project. The presentation was given on 11th April 2014.]]> Sat, 12 Apr 2014 14:50:45 GMT /slideshow/utilizing-additional-information-in-factorization-methods-research-overview-april-2014/33450437 balazshidasi@slideshare.net(balazshidasi) Utilizing additional information in factorization methods (research overview, April 2014) balazshidasi This presentation contains the main points of my recommender systems related research. It describes the arc of my research starting from improving matrix factorization, through the developement of my context-aware algorithms & addressing scalability issues to developing a general factorization framework & dealing with context dimension modeling. The slides were presented at the Delft University of Technology where I was invited to give this introductory talk as part of the collaboration between participiants of the CrowdRec project. The presentation was given on 11th April 2014. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/tudprezi-140412145045-phpapp01-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> This presentation contains the main points of my recommender systems related research. It describes the arc of my research starting from improving matrix factorization, through the developement of my context-aware algorithms &amp; addressing scalability issues to developing a general factorization framework &amp; dealing with context dimension modeling. The slides were presented at the Delft University of Technology where I was invited to give this introductory talk as part of the collaboration between participiants of the CrowdRec project. The presentation was given on 11th April 2014.
Utilizing additional information in factorization methods (research overview, April 2014) from Bal叩zs Hidasi
]]> 628 7 https://cdn.slidesharecdn.com/ss_thumbnails/tudprezi-140412145045-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation 000000 http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Az implicit ajánlási probléma és néhány megoldása (BME TMIT szeminárium előadás, 2012) /slideshow/azimplicitajnlsi-problma-s-nhny-megoldsa-bme-tmit-szeminrium-elads-2012/16464387 recocontexttmitseminar-130211043239-phpapp02
Ez a diasor egy ismeretterjesztő előadáshoz készült. Az előadás témája az implicit feedback alapú ajánlás (amikor a felhasználók preferenciái nem olvashatóak ki közvetlenül az adatokból), és a probléma néhány lehetséges megoldása. A prezentáció a probléma ismertetését követően kitér néhány kutatási eredményemre, mint például a mátrix faktorizáció inicializálására, vagy az implicit tenzorfaktorizációra. Az előadásra 2012. nyarán, a BME Távközlési és Médiainformatikai Tanszéke (TMIT) által szervezett szemináriumon került sor.]]>
Ez a diasor egy ismeretterjesztő előadáshoz készült. Az előadás témája az implicit feedback alapú ajánlás (amikor a felhasználók preferenciái nem olvashatóak ki közvetlenül az adatokból), és a probléma néhány lehetséges megoldása. A prezentáció a probléma ismertetését követően kitér néhány kutatási eredményemre, mint például a mátrix faktorizáció inicializálására, vagy az implicit tenzorfaktorizációra. Az előadásra 2012. nyarán, a BME Távközlési és Médiainformatikai Tanszéke (TMIT) által szervezett szemináriumon került sor.]]> Mon, 11 Feb 2013 04:32:39 GMT /slideshow/azimplicitajnlsi-problma-s-nhny-megoldsa-bme-tmit-szeminrium-elads-2012/16464387 balazshidasi@slideshare.net(balazshidasi) Az implicit ajánlási probléma és néhány megoldása (BME TMIT szeminárium előadás, 2012) balazshidasi Ez a diasor egy ismeretterjesztő előadáshoz készült. Az előadás témája az implicit feedback alapú ajánlás (amikor a felhasználók preferenciái nem olvashatóak ki közvetlenül az adatokból), és a probléma néhány lehetséges megoldása. A prezentáció a probléma ismertetését követően kitér néhány kutatási eredményemre, mint például a mátrix faktorizáció inicializálására, vagy az implicit tenzorfaktorizációra. Az előadásra 2012. nyarán, a BME Távközlési és Médiainformatikai Tanszéke (TMIT) által szervezett szemináriumon került sor. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/recocontexttmitseminar-130211043239-phpapp02-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Ez a diasor egy ismeretterjesztő előadáshoz készült. Az előadás témája az implicit feedback alapú ajánlás (amikor a felhasználók preferenciái nem olvashatóak ki közvetlenül az adatokból), és a probléma néhány lehetséges megoldása. A prezentáció a probléma ismertetését követően kitér néhány kutatási eredményemre, mint például a mátrix faktorizáció inicializálására, vagy az implicit tenzorfaktorizációra. Az előadásra 2012. nyarán, a BME Távközlési és Médiainformatikai Tanszéke (TMIT) által szervezett szemináriumon került sor.
Az implicit ajテ。nlテ。si problテゥma テゥs nテゥhテ。ny megoldテ。sa (BME TMIT szeminテ。rium elナ疎dテ。s, 2012) from Balテ。zs Hidasi
]]> 383 2 https://cdn.slidesharecdn.com/ss_thumbnails/recocontexttmitseminar-130211043239-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Context-aware similarities within the factorization framework (CaRR 2013 presentation) /slideshow/contextaware-similarities-within-the-factorization-framework-carr-2013-presentation/16464057 casimcarr13-130211040838-phpapp01
This presentation is about an interesting side project of my main research in recommender systems. It is about the preliminary examination of context-aware similarities in the factorization framework. This work is in the intersection of the following areas: (1) implicit feedback based recommendations; (2) context / context awareness; (3) item-to-item recommendations; (4) matrix / tensor factorization. The aim of this work is to examine whether context can be used to compute more accurate item similarities based on their feature vectors. Two levels of context aware similarities are introduced: (1) context is only used during training, but not for computing the similarity; (2) context is used during the training and for the similarity computations as well. This presentation was given at the 3rd workshop on Context-awareness in Retrieval and Recommendations (CaRR 2013) in Rome.]]>
This presentation is about an interesting side project of my main research in recommender systems. It is about the preliminary examination of context-aware similarities in the factorization framework. This work is in the intersection of the following areas: (1) implicit feedback based recommendations; (2) context / context awareness; (3) item-to-item recommendations; (4) matrix / tensor factorization. The aim of this work is to examine whether context can be used to compute more accurate item similarities based on their feature vectors. Two levels of context aware similarities are introduced: (1) context is only used during training, but not for computing the similarity; (2) context is used during the training and for the similarity computations as well. This presentation was given at the 3rd workshop on Context-awareness in Retrieval and Recommendations (CaRR 2013) in Rome.]]> Mon, 11 Feb 2013 04:08:37 GMT /slideshow/contextaware-similarities-within-the-factorization-framework-carr-2013-presentation/16464057 balazshidasi@slideshare.net(balazshidasi) Context-aware similarities within the factorization framework (CaRR 2013 presentation) balazshidasi This presentation is about an interesting side project of my main research in recommender systems. It is about the preliminary examination of context-aware similarities in the factorization framework. This work is in the intersection of the following areas: (1) implicit feedback based recommendations; (2) context / context awareness; (3) item-to-item recommendations; (4) matrix / tensor factorization. The aim of this work is to examine whether context can be used to compute more accurate item similarities based on their feature vectors. Two levels of context aware similarities are introduced: (1) context is only used during training, but not for computing the similarity; (2) context is used during the training and for the similarity computations as well. This presentation was given at the 3rd workshop on Context-awareness in Retrieval and Recommendations (CaRR 2013) in Rome. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/casimcarr13-130211040838-phpapp01-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> This presentation is about an interesting side project of my main research in recommender systems. It is about the preliminary examination of context-aware similarities in the factorization framework. This work is in the intersection of the following areas: (1) implicit feedback based recommendations; (2) context / context awareness; (3) item-to-item recommendations; (4) matrix / tensor factorization. The aim of this work is to examine whether context can be used to compute more accurate item similarities based on their feature vectors. Two levels of context aware similarities are introduced: (1) context is only used during training, but not for computing the similarity; (2) context is used during the training and for the similarity computations as well. This presentation was given at the 3rd workshop on Context-awareness in Retrieval and Recommendations (CaRR 2013) in Rome.
Context-aware similarities within the factorization framework (CaRR 2013 presentation) from Bal叩zs Hidasi
]]> 1194 2 https://cdn.slidesharecdn.com/ss_thumbnails/casimcarr13-130211040838-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 iTALS: implicit tensor factorization for context-aware recommendations (ECML/PKDD 2012 presentation) /slideshow/itals-implicit-tensor-factorization-for-contextaware-recommendations-ecmlpkdd-2012-presentation/16463953 italsecml12-130211040125-phpapp02
This presentation is about the context-aware recommender algorithm iTALS. iTALS is a context-aware recommender algorithm for implicit feedback data. The user-item-context(s) setup is modelled in a binary tensor. Weights are also assigned to the cells based on the certainity of their information. An ALS-based algorithm is proposed that is capable of efficiently factorizing this tensor. Additionally a novel context information is introduced: sequentiality. This context allows us to incorporate association rule like information into the factorization framework and to differentiate between items with different repetetiveness patters and thus to make recommendations more accurate. This presentation was originally given at ECML/PKDD 2012 in Bristol.]]>
This presentation is about the context-aware recommender algorithm iTALS. iTALS is a context-aware recommender algorithm for implicit feedback data. The user-item-context(s) setup is modelled in a binary tensor. Weights are also assigned to the cells based on the certainity of their information. An ALS-based algorithm is proposed that is capable of efficiently factorizing this tensor. Additionally a novel context information is introduced: sequentiality. This context allows us to incorporate association rule like information into the factorization framework and to differentiate between items with different repetetiveness patters and thus to make recommendations more accurate. This presentation was originally given at ECML/PKDD 2012 in Bristol.]]> Mon, 11 Feb 2013 04:01:25 GMT /slideshow/itals-implicit-tensor-factorization-for-contextaware-recommendations-ecmlpkdd-2012-presentation/16463953 balazshidasi@slideshare.net(balazshidasi) iTALS: implicit tensor factorization for context-aware recommendations (ECML/PKDD 2012 presentation) balazshidasi This presentation is about the context-aware recommender algorithm iTALS. iTALS is a context-aware recommender algorithm for implicit feedback data. The user-item-context(s) setup is modelled in a binary tensor. Weights are also assigned to the cells based on the certainity of their information. An ALS-based algorithm is proposed that is capable of efficiently factorizing this tensor. Additionally a novel context information is introduced: sequentiality. This context allows us to incorporate association rule like information into the factorization framework and to differentiate between items with different repetetiveness patters and thus to make recommendations more accurate. This presentation was originally given at ECML/PKDD 2012 in Bristol. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/italsecml12-130211040125-phpapp02-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> This presentation is about the context-aware recommender algorithm iTALS. iTALS is a context-aware recommender algorithm for implicit feedback data. The user-item-context(s) setup is modelled in a binary tensor. Weights are also assigned to the cells based on the certainity of their information. An ALS-based algorithm is proposed that is capable of efficiently factorizing this tensor. Additionally a novel context information is introduced: sequentiality. This context allows us to incorporate association rule like information into the factorization framework and to differentiate between items with different repetetiveness patters and thus to make recommendations more accurate. This presentation was originally given at ECML/PKDD 2012 in Bristol.
iTALS: implicit tensor factorization for context-aware recommendations (ECML/PKDD 2012 presentation) from Bal叩zs Hidasi
]]> 638 4 https://cdn.slidesharecdn.com/ss_thumbnails/italsecml12-130211040125-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 Initialization of matrix factorization (CaRR 2012 presentation) /slideshow/initialization-of-matrix-factorization-carr-2012-presentation/16457700 mfinitcarr12-130210163646-phpapp01
This presentation is about why initialization of matrix factorization methods is important and proposes an interesting initialization method (coined SimFactor). The method revolves around a similarity preserving dimensionality reduction technique. Context-based initialization is introduced as well. As most of my recommender systems related research, this presentation focuses on implicit feedback (the case where user preferences are not coded explicitely in the data). Originally presented at the 2nd workshop on Context-awareness in Retrieval and Recommendations (CaRR 2012) in Lisbon.]]>
This presentation is about why initialization of matrix factorization methods is important and proposes an interesting initialization method (coined SimFactor). The method revolves around a similarity preserving dimensionality reduction technique. Context-based initialization is introduced as well. As most of my recommender systems related research, this presentation focuses on implicit feedback (the case where user preferences are not coded explicitely in the data). Originally presented at the 2nd workshop on Context-awareness in Retrieval and Recommendations (CaRR 2012) in Lisbon.]]> Sun, 10 Feb 2013 16:36:46 GMT /slideshow/initialization-of-matrix-factorization-carr-2012-presentation/16457700 balazshidasi@slideshare.net(balazshidasi) Initialization of matrix factorization (CaRR 2012 presentation) balazshidasi This presentation is about why initialization of matrix factorization methods is important and proposes an interesting initialization method (coined SimFactor). The method revolves around a similarity preserving dimensionality reduction technique. Context-based initialization is introduced as well. As most of my recommender systems related research, this presentation focuses on implicit feedback (the case where user preferences are not coded explicitely in the data). Originally presented at the 2nd workshop on Context-awareness in Retrieval and Recommendations (CaRR 2012) in Lisbon. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/mfinitcarr12-130210163646-phpapp01-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> This presentation is about why initialization of matrix factorization methods is important and proposes an interesting initialization method (coined SimFactor). The method revolves around a similarity preserving dimensionality reduction technique. Context-based initialization is introduced as well. As most of my recommender systems related research, this presentation focuses on implicit feedback (the case where user preferences are not coded explicitely in the data). Originally presented at the 2nd workshop on Context-awareness in Retrieval and Recommendations (CaRR 2012) in Lisbon.
Initialization of matrix factorization (CaRR 2012 presentation) from Bal叩zs Hidasi
]]> 737 3 https://cdn.slidesharecdn.com/ss_thumbnails/mfinitcarr12-130210163646-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 ShiftTree: model alapú idősor-osztályozó (VK 2009 előadás) /slideshow/shifttree-model-alap-idsorosztlyoz-vk-2009-elads/16456882 shifttreevknoanim-130210145717-phpapp01
A prezentáció témája a ShiftTree névre hallgató, egyedi, model alapú idősor-osztályozó. A ShiftTree az idősor-osztályozás problémájának egy egyedülálló, modell alapú megközelítése. Az elképzelés alapja, hogy minden idősorhoz egy szemet (kurzort) rendelünk, ami az időtengely egy adott pontjára mutat. Dinamikus attribútumokat hozunk létre úgy, hogy a következő két kérdésre válaszolunk: (1) Hová nézzünk az időtengelyen? (2) Mit nézzünk az adott pontban? Az első kérdésre adott válasz azt mondja meg, hogy hogyan mozgassuk a szemet az időtengely mentén. A második válasz pedig azt definiálja, hogy hogyan számoljuk ki a dinamikus attribútum értékét az adott pontban. Ezeket a dinamikus attribútumokat ezután egy bináris döntési fában használjuk fel. Ez a diasor a ShiftTree egy korai (2009-es) verzióját mutatja be. A prezentáció a 2009-es Végzős Konferencián került bemutatásra. Megjegyzés: valamilyen oknál fogva a ݺߣShare nem támogatja az animációkat, ezért az animált diák több diára lettek szétszedve.]]>
A prezentáció témája a ShiftTree névre hallgató, egyedi, model alapú idősor-osztályozó. A ShiftTree az idősor-osztályozás problémájának egy egyedülálló, modell alapú megközelítése. Az elképzelés alapja, hogy minden idősorhoz egy szemet (kurzort) rendelünk, ami az időtengely egy adott pontjára mutat. Dinamikus attribútumokat hozunk létre úgy, hogy a következő két kérdésre válaszolunk: (1) Hová nézzünk az időtengelyen? (2) Mit nézzünk az adott pontban? Az első kérdésre adott válasz azt mondja meg, hogy hogyan mozgassuk a szemet az időtengely mentén. A második válasz pedig azt definiálja, hogy hogyan számoljuk ki a dinamikus attribútum értékét az adott pontban. Ezeket a dinamikus attribútumokat ezután egy bináris döntési fában használjuk fel. Ez a diasor a ShiftTree egy korai (2009-es) verzióját mutatja be. A prezentáció a 2009-es Végzős Konferencián került bemutatásra. Megjegyzés: valamilyen oknál fogva a ݺߣShare nem támogatja az animációkat, ezért az animált diák több diára lettek szétszedve.]]> Sun, 10 Feb 2013 14:57:17 GMT /slideshow/shifttree-model-alap-idsorosztlyoz-vk-2009-elads/16456882 balazshidasi@slideshare.net(balazshidasi) ShiftTree: model alapú idősor-osztályozó (VK 2009 előadás) balazshidasi A prezentáció témája a ShiftTree névre hallgató, egyedi, model alapú idősor-osztályozó. A ShiftTree az idősor-osztályozás problémájának egy egyedülálló, modell alapú megközelítése. Az elképzelés alapja, hogy minden idősorhoz egy szemet (kurzort) rendelünk, ami az időtengely egy adott pontjára mutat. Dinamikus attribútumokat hozunk létre úgy, hogy a következő két kérdésre válaszolunk: (1) Hová nézzünk az időtengelyen? (2) Mit nézzünk az adott pontban? Az első kérdésre adott válasz azt mondja meg, hogy hogyan mozgassuk a szemet az időtengely mentén. A második válasz pedig azt definiálja, hogy hogyan számoljuk ki a dinamikus attribútum értékét az adott pontban. Ezeket a dinamikus attribútumokat ezután egy bináris döntési fában használjuk fel. Ez a diasor a ShiftTree egy korai (2009-es) verzióját mutatja be. A prezentáció a 2009-es Végzős Konferencián került bemutatásra. Megjegyzés: valamilyen oknál fogva a ݺߣShare nem támogatja az animációkat, ezért az animált diák több diára lettek szétszedve. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/shifttreevknoanim-130210145717-phpapp01-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> A prezentáció témája a ShiftTree névre hallgató, egyedi, model alapú idősor-osztályozó. A ShiftTree az idősor-osztályozás problémájának egy egyedülálló, modell alapú megközelítése. Az elképzelés alapja, hogy minden idősorhoz egy szemet (kurzort) rendelünk, ami az időtengely egy adott pontjára mutat. Dinamikus attribútumokat hozunk létre úgy, hogy a következő két kérdésre válaszolunk: (1) Hová nézzünk az időtengelyen? (2) Mit nézzünk az adott pontban? Az első kérdésre adott válasz azt mondja meg, hogy hogyan mozgassuk a szemet az időtengely mentén. A második válasz pedig azt definiálja, hogy hogyan számoljuk ki a dinamikus attribútum értékét az adott pontban. Ezeket a dinamikus attribútumokat ezután egy bináris döntési fában használjuk fel. Ez a diasor a ShiftTree egy korai (2009-es) verzióját mutatja be. A prezentáció a 2009-es Végzős Konferencián került bemutatásra. Megjegyzés: valamilyen oknál fogva a ݺߣShare nem támogatja az animációkat, ezért az animált diák több diára lettek szétszedve.
ShiftTree: model alapテコ idナ壮or-osztテ。lyozテウ (VK 2009 elナ疎dテ。s) from Balテ。zs Hidasi
]]> 328 2 https://cdn.slidesharecdn.com/ss_thumbnails/shifttreevknoanim-130210145717-phpapp01-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 ShiftTree: model alapú idősor-osztályozó (ML@BP előadás, 2012) /slideshow/shifttree-mlbp-noanim/16456821 shifttreemlbpnoanim-130210144748-phpapp02
A prezentáció témája a ShiftTree névre hallgató, egyedi, model alapú idősor-osztályozó. A ShiftTree az idősor-osztályozás problémájának egy egyedülálló, modell alapú megközelítése. Az elképzelés alapja, hogy minden idősorhoz egy szemet (kurzort) rendelünk, ami az időtengely egy adott pontjára mutat. Dinamikus attribútumokat hozunk létre úgy, hogy a következő két kérdésre válaszolunk: (1) Hová nézzünk az időtengelyen? (2) Mit nézzünk az adott pontban? Az első kérdésre adott válasz azt mondja meg, hogy hogyan mozgassuk a szemet az időtengely mentén. A második válasz pedig azt definiálja, hogy hogyan számoljuk ki a dinamikus attribútum értékét az adott pontban. Ezeket a dinamikus attribútumokat ezután egy bináris döntési fában használjuk fel. Ez a diasor a legteljesebb, a ShiftTree-ről szóló prezentációk közül. Tartalmaz több kiegészítést, valamint leír néhány olyan megoldást, amik a kutatás során előkerültek, de végül zsákutcának bizonyultak. A prezentáció egy 2012. februári előadáshoz tartozik, amire az ML@BP rendezvénysorozat keretein belül került sor. Megjegyzés: valamilyen oknál fogva a ݺߣShare nem támogatja az animációkat, ezért az animált diák több diára lettek szétszedve.]]>
A prezentáció témája a ShiftTree névre hallgató, egyedi, model alapú idősor-osztályozó. A ShiftTree az idősor-osztályozás problémájának egy egyedülálló, modell alapú megközelítése. Az elképzelés alapja, hogy minden idősorhoz egy szemet (kurzort) rendelünk, ami az időtengely egy adott pontjára mutat. Dinamikus attribútumokat hozunk létre úgy, hogy a következő két kérdésre válaszolunk: (1) Hová nézzünk az időtengelyen? (2) Mit nézzünk az adott pontban? Az első kérdésre adott válasz azt mondja meg, hogy hogyan mozgassuk a szemet az időtengely mentén. A második válasz pedig azt definiálja, hogy hogyan számoljuk ki a dinamikus attribútum értékét az adott pontban. Ezeket a dinamikus attribútumokat ezután egy bináris döntési fában használjuk fel. Ez a diasor a legteljesebb, a ShiftTree-ről szóló prezentációk közül. Tartalmaz több kiegészítést, valamint leír néhány olyan megoldást, amik a kutatás során előkerültek, de végül zsákutcának bizonyultak. A prezentáció egy 2012. februári előadáshoz tartozik, amire az ML@BP rendezvénysorozat keretein belül került sor. Megjegyzés: valamilyen oknál fogva a ݺߣShare nem támogatja az animációkat, ezért az animált diák több diára lettek szétszedve.]]> Sun, 10 Feb 2013 14:47:48 GMT /slideshow/shifttree-mlbp-noanim/16456821 balazshidasi@slideshare.net(balazshidasi) ShiftTree: model alapú idősor-osztályozó (ML@BP előadás, 2012) balazshidasi A prezentáció témája a ShiftTree névre hallgató, egyedi, model alapú idősor-osztályozó. A ShiftTree az idősor-osztályozás problémájának egy egyedülálló, modell alapú megközelítése. Az elképzelés alapja, hogy minden idősorhoz egy szemet (kurzort) rendelünk, ami az időtengely egy adott pontjára mutat. Dinamikus attribútumokat hozunk létre úgy, hogy a következő két kérdésre válaszolunk: (1) Hová nézzünk az időtengelyen? (2) Mit nézzünk az adott pontban? Az első kérdésre adott válasz azt mondja meg, hogy hogyan mozgassuk a szemet az időtengely mentén. A második válasz pedig azt definiálja, hogy hogyan számoljuk ki a dinamikus attribútum értékét az adott pontban. Ezeket a dinamikus attribútumokat ezután egy bináris döntési fában használjuk fel. Ez a diasor a legteljesebb, a ShiftTree-ről szóló prezentációk közül. Tartalmaz több kiegészítést, valamint leír néhány olyan megoldást, amik a kutatás során előkerültek, de végül zsákutcának bizonyultak. A prezentáció egy 2012. februári előadáshoz tartozik, amire az ML@BP rendezvénysorozat keretein belül került sor. Megjegyzés: valamilyen oknál fogva a ݺߣShare nem támogatja az animációkat, ezért az animált diák több diára lettek szétszedve. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/shifttreemlbpnoanim-130210144748-phpapp02-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> A prezentáció témája a ShiftTree névre hallgató, egyedi, model alapú idősor-osztályozó. A ShiftTree az idősor-osztályozás problémájának egy egyedülálló, modell alapú megközelítése. Az elképzelés alapja, hogy minden idősorhoz egy szemet (kurzort) rendelünk, ami az időtengely egy adott pontjára mutat. Dinamikus attribútumokat hozunk létre úgy, hogy a következő két kérdésre válaszolunk: (1) Hová nézzünk az időtengelyen? (2) Mit nézzünk az adott pontban? Az első kérdésre adott válasz azt mondja meg, hogy hogyan mozgassuk a szemet az időtengely mentén. A második válasz pedig azt definiálja, hogy hogyan számoljuk ki a dinamikus attribútum értékét az adott pontban. Ezeket a dinamikus attribútumokat ezután egy bináris döntési fában használjuk fel. Ez a diasor a legteljesebb, a ShiftTree-ről szóló prezentációk közül. Tartalmaz több kiegészítést, valamint leír néhány olyan megoldást, amik a kutatás során előkerültek, de végül zsákutcának bizonyultak. A prezentáció egy 2012. februári előadáshoz tartozik, amire az ML@BP rendezvénysorozat keretein belül került sor. Megjegyzés: valamilyen oknál fogva a ݺߣShare nem támogatja az animációkat, ezért az animált diák több diára lettek szétszedve.
ShiftTree: model alapテコ idナ壮or-osztテ。lyozテウ (ML@BP elナ疎dテ。s, 2012) from Balテ。zs Hidasi
]]> 358 2 https://cdn.slidesharecdn.com/ss_thumbnails/shifttreemlbpnoanim-130210144748-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 ShiftTree: model based time series classifier (ECML/PKDD 2011 presentation) /slideshow/shifttree-ecml11/16380410 shifttreeecml11-130206072811-phpapp02
This slideshow is about the time series classifier algorithm, ShiftTree. ShiftTree is a unique, model-based approach for time series classification. The basic idea is that we assign a cursor (or eye) to each series and move this to certain positions on the time axis. We generate dynamic attributes by answering two questions: (1) Where to look? (2) What to look at?. The answer to the first question tells us where to move the cursor (e.g.: forward 100 steps, to the previous local maxima, etc), while the second answer defines the calculation of the dynamic attributes (e.g.: value at that point, the weighted avarage of the values around the position, the difference in the current and previous cursor position, etc). These dynamic attributes then used in a binary decision tree. This slideshow was originally presented at ECML/PKDD 2011 in Athens. Note that for whatever reasons ݺߣShare doesn't support animations. Therefore the animated slides were split into multiple slides.]]>
This slideshow is about the time series classifier algorithm, ShiftTree. ShiftTree is a unique, model-based approach for time series classification. The basic idea is that we assign a cursor (or eye) to each series and move this to certain positions on the time axis. We generate dynamic attributes by answering two questions: (1) Where to look? (2) What to look at?. The answer to the first question tells us where to move the cursor (e.g.: forward 100 steps, to the previous local maxima, etc), while the second answer defines the calculation of the dynamic attributes (e.g.: value at that point, the weighted avarage of the values around the position, the difference in the current and previous cursor position, etc). These dynamic attributes then used in a binary decision tree. This slideshow was originally presented at ECML/PKDD 2011 in Athens. Note that for whatever reasons ݺߣShare doesn't support animations. Therefore the animated slides were split into multiple slides.]]> Wed, 06 Feb 2013 07:28:11 GMT /slideshow/shifttree-ecml11/16380410 balazshidasi@slideshare.net(balazshidasi) ShiftTree: model based time series classifier (ECML/PKDD 2011 presentation) balazshidasi This slideshow is about the time series classifier algorithm, ShiftTree. ShiftTree is a unique, model-based approach for time series classification. The basic idea is that we assign a cursor (or eye) to each series and move this to certain positions on the time axis. We generate dynamic attributes by answering two questions: (1) Where to look? (2) What to look at?. The answer to the first question tells us where to move the cursor (e.g.: forward 100 steps, to the previous local maxima, etc), while the second answer defines the calculation of the dynamic attributes (e.g.: value at that point, the weighted avarage of the values around the position, the difference in the current and previous cursor position, etc). These dynamic attributes then used in a binary decision tree. This slideshow was originally presented at ECML/PKDD 2011 in Athens. Note that for whatever reasons ݺߣShare doesn't support animations. Therefore the animated slides were split into multiple slides. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/shifttreeecml11-130206072811-phpapp02-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> This slideshow is about the time series classifier algorithm, ShiftTree. ShiftTree is a unique, model-based approach for time series classification. The basic idea is that we assign a cursor (or eye) to each series and move this to certain positions on the time axis. We generate dynamic attributes by answering two questions: (1) Where to look? (2) What to look at?. The answer to the first question tells us where to move the cursor (e.g.: forward 100 steps, to the previous local maxima, etc), while the second answer defines the calculation of the dynamic attributes (e.g.: value at that point, the weighted avarage of the values around the position, the difference in the current and previous cursor position, etc). These dynamic attributes then used in a binary decision tree. This slideshow was originally presented at ECML/PKDD 2011 in Athens. Note that for whatever reasons ݺߣShare doesn&#39;t support animations. Therefore the animated slides were split into multiple slides.
ShiftTree: model based time series classifier (ECML/PKDD 2011 presentation) from Bal叩zs Hidasi
]]> 731 4 https://cdn.slidesharecdn.com/ss_thumbnails/shifttreeecml11-130206072811-phpapp02-thumbnail.jpg?width=120&height=120&fit=bounds presentation White http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 https://cdn.slidesharecdn.com/profile-photo-balazshidasi-48x48.jpg?cb=1700740752 I'm a machine learning researcher with a wide interest in the field of machine learning and data mining. My current focus topics are deep learning, recommender systems and collaborative filtering; earlier I worked on model based time series classification. Currently I am the Head of Data Mining and Research at Gravity R&D, a Budapest (Hungary) based recommendation service provider company. I lead a small team of data scientists and work on my own recommendation systems related research. I actively publish the results of my research at scientific conferences and journals. I am an active participant in the EU FP7 funded CrowdRec project that aims at developing the new generation of recomm... www.hidasi.eu https://cdn.slidesharecdn.com/ss_thumbnails/genaimeetup20231114-231115135405-b9a22d53-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/egyedi-termk-kreatvok-tmeges-gyrtsa-generatv-ai-segtsgvel/263459147 Egyedi termék kreatívo... https://cdn.slidesharecdn.com/ss_thumbnails/thirdpartyevalsplitslides-230920031211-dbcfe40c-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/the-effect-of-third-party-implementations-on-reproducibility/261174582 The Effect of Third Pa... https://cdn.slidesharecdn.com/ss_thumbnails/gru4recv2cikm18-181024131645-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/gru4rec-v2-recurrent-neural-networks-with-topk-gains-for-sessionbased-recommendations/120561602 GRU4Rec v2 - Recurrent...