home.social

#optimaltransport — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #optimaltransport, aggregated by home.social.

fetched live
  1. Following up on this, I also explored a more direct use of #WassersteinDistance in #WGANs: Instead of training a discriminator, the generator is optimized by explicitly computing the #OptimalTransport distance between real and generated samples. This turns the loss into the actual metric of interest and removes the adversarial setup, leading to a more direct and stable training signal. And we can generate cool animations, too ^_^

    🌍 fabriziomusacchio.com/blog/202

    #MachineLearning #Wasserstein

  2. Following up on this, I also explored a more direct use of #WassersteinDistance in #WGANs: Instead of training a discriminator, the generator is optimized by explicitly computing the #OptimalTransport distance between real and generated samples. This turns the loss into the actual metric of interest and removes the adversarial setup, leading to a more direct and stable training signal. And we can generate cool animations, too ^_^

    🌍 fabriziomusacchio.com/blog/202

    #MachineLearning #Wasserstein

  3. Following up on this, I also explored a more direct use of #WassersteinDistance in #WGANs: Instead of training a discriminator, the generator is optimized by explicitly computing the #OptimalTransport distance between real and generated samples. This turns the loss into the actual metric of interest and removes the adversarial setup, leading to a more direct and stable training signal. And we can generate cool animations, too ^_^

    🌍 fabriziomusacchio.com/blog/202

    #MachineLearning #Wasserstein

  4. Following up on this, I also explored a more direct use of #WassersteinDistance in #WGANs: Instead of training a discriminator, the generator is optimized by explicitly computing the #OptimalTransport distance between real and generated samples. This turns the loss into the actual metric of interest and removes the adversarial setup, leading to a more direct and stable training signal. And we can generate cool animations, too ^_^

    🌍 fabriziomusacchio.com/blog/202

    #MachineLearning #Wasserstein

  5. Following up on this, I also explored a more direct use of #WassersteinDistance in #WGANs: Instead of training a discriminator, the generator is optimized by explicitly computing the #OptimalTransport distance between real and generated samples. This turns the loss into the actual metric of interest and removes the adversarial setup, leading to a more direct and stable training signal. And we can generate cool animations, too ^_^

    🌍 fabriziomusacchio.com/blog/202

    #MachineLearning #Wasserstein

  6. 📐📚New study on #WassersteinDistance: Bonet et al. study #geodesic rays in #Wasserstein space and derive conditions for their existence. They show that #Busemann functions can be computed via #OT, with closed-form solutions for 1D and Gaussian cases. This enables efficient sliced distances for labeled datasets, closely matching classical metrics at lower cost and supporting dataset “flows” for #TransferLearning.

    🌍 openreview.net/forum?id=Xpt0HE

    #OptimalTransport #MachineLearning

  7. 📐📚New study on #WassersteinDistance: Bonet et al. study #geodesic rays in #Wasserstein space and derive conditions for their existence. They show that #Busemann functions can be computed via #OT, with closed-form solutions for 1D and Gaussian cases. This enables efficient sliced distances for labeled datasets, closely matching classical metrics at lower cost and supporting dataset “flows” for #TransferLearning.

    🌍 openreview.net/forum?id=Xpt0HE

    #OptimalTransport #MachineLearning

  8. 📐📚New study on #WassersteinDistance: Bonet et al. study #geodesic rays in #Wasserstein space and derive conditions for their existence. They show that #Busemann functions can be computed via #OT, with closed-form solutions for 1D and Gaussian cases. This enables efficient sliced distances for labeled datasets, closely matching classical metrics at lower cost and supporting dataset “flows” for #TransferLearning.

    🌍 openreview.net/forum?id=Xpt0HE

    #OptimalTransport #MachineLearning

  9. 📐📚New study on #WassersteinDistance: Bonet et al. study #geodesic rays in #Wasserstein space and derive conditions for their existence. They show that #Busemann functions can be computed via #OT, with closed-form solutions for 1D and Gaussian cases. This enables efficient sliced distances for labeled datasets, closely matching classical metrics at lower cost and supporting dataset “flows” for #TransferLearning.

    🌍 openreview.net/forum?id=Xpt0HE

    #OptimalTransport #MachineLearning

  10. 📐📚New study on #WassersteinDistance: Bonet et al. study #geodesic rays in #Wasserstein space and derive conditions for their existence. They show that #Busemann functions can be computed via #OT, with closed-form solutions for 1D and Gaussian cases. This enables efficient sliced distances for labeled datasets, closely matching classical metrics at lower cost and supporting dataset “flows” for #TransferLearning.

    🌍 openreview.net/forum?id=Xpt0HE

    #OptimalTransport #MachineLearning

  11. 📐 New preprint by Gabriel Peyré: The paper introduces a new class of spectral #Wasserstein distances, linking #OptimalTransport with normalized #gradient methods. It shows that spectrally normalized #GradientDescent can be interpreted as a gradient flow in this spectral-W geometry, providing a principled bridge between #optimization dynamics and transport metrics:

    📄 arxiv.org/abs/2604.04891

    #MachineLearning #WassersteinDistance

  12. 📐 New preprint by Gabriel Peyré: The paper introduces a new class of spectral #Wasserstein distances, linking #OptimalTransport with normalized #gradient methods. It shows that spectrally normalized #GradientDescent can be interpreted as a gradient flow in this spectral-W geometry, providing a principled bridge between #optimization dynamics and transport metrics:

    📄 arxiv.org/abs/2604.04891

    #MachineLearning #WassersteinDistance

  13. 📐 New preprint by Gabriel Peyré: The paper introduces a new class of spectral #Wasserstein distances, linking #OptimalTransport with normalized #gradient methods. It shows that spectrally normalized #GradientDescent can be interpreted as a gradient flow in this spectral-W geometry, providing a principled bridge between #optimization dynamics and transport metrics:

    📄 arxiv.org/abs/2604.04891

    #MachineLearning #WassersteinDistance

  14. 📐 New preprint by Gabriel Peyré: The paper introduces a new class of spectral #Wasserstein distances, linking #OptimalTransport with normalized #gradient methods. It shows that spectrally normalized #GradientDescent can be interpreted as a gradient flow in this spectral-W geometry, providing a principled bridge between #optimization dynamics and transport metrics:

    📄 arxiv.org/abs/2604.04891

    #MachineLearning #WassersteinDistance

  15. 📐 New preprint by Gabriel Peyré: The paper introduces a new class of spectral #Wasserstein distances, linking #OptimalTransport with normalized #gradient methods. It shows that spectrally normalized #GradientDescent can be interpreted as a gradient flow in this spectral-W geometry, providing a principled bridge between #optimization dynamics and transport metrics:

    📄 arxiv.org/abs/2604.04891

    #MachineLearning #WassersteinDistance

  16. 📝💤 "Behold, the 'brief' intro to optimal transport where intuition triumphs over 'maths' because who needs rigor? 🙄 It's basically a #YouTube rabbit hole disguised as a blog, because nothing says 'understandable' like suggesting you watch a four-year-old lecture series. 📚📺"
    alexhwilliams.info/itsneuronal #optimaltransport #rabbitHole #blogpost #mathintuition #lectureSeries #HackerNews #ngated

  17. 📝💤 "Behold, the 'brief' intro to optimal transport where intuition triumphs over 'maths' because who needs rigor? 🙄 It's basically a #YouTube rabbit hole disguised as a blog, because nothing says 'understandable' like suggesting you watch a four-year-old lecture series. 📚📺"
    alexhwilliams.info/itsneuronal #optimaltransport #rabbitHole #blogpost #mathintuition #lectureSeries #HackerNews #ngated

  18. 📝💤 "Behold, the 'brief' intro to optimal transport where intuition triumphs over 'maths' because who needs rigor? 🙄 It's basically a #YouTube rabbit hole disguised as a blog, because nothing says 'understandable' like suggesting you watch a four-year-old lecture series. 📚📺"
    alexhwilliams.info/itsneuronal #optimaltransport #rabbitHole #blogpost #mathintuition #lectureSeries #HackerNews #ngated

  19. 📝💤 "Behold, the 'brief' intro to optimal transport where intuition triumphs over 'maths' because who needs rigor? 🙄 It's basically a #YouTube rabbit hole disguised as a blog, because nothing says 'understandable' like suggesting you watch a four-year-old lecture series. 📚📺"
    alexhwilliams.info/itsneuronal #optimaltransport #rabbitHole #blogpost #mathintuition #lectureSeries #HackerNews #ngated

  20. 📝💤 "Behold, the 'brief' intro to optimal transport where intuition triumphs over 'maths' because who needs rigor? 🙄 It's basically a #YouTube rabbit hole disguised as a blog, because nothing says 'understandable' like suggesting you watch a four-year-old lecture series. 📚📺"
    alexhwilliams.info/itsneuronal #optimaltransport #rabbitHole #blogpost #mathintuition #lectureSeries #HackerNews #ngated

  21. The #Wasserstein distance (#EMD), sliced Wasserstein distance (#SWD), and the #L2norm are common #metrics used to quantify the ‘distance’ between two distributions. This tutorial compares these three metrics and discusses their advantages and disadvantages.

    🌎 fabriziomusacchio.com/blog/202

    #OptimalTransport #MachineLearning

  22. The #Wasserstein distance (#EMD), sliced Wasserstein distance (#SWD), and the #L2norm are common #metrics used to quantify the ‘distance’ between two distributions. This tutorial compares these three metrics and discusses their advantages and disadvantages.

    🌎 fabriziomusacchio.com/blog/202

    #OptimalTransport #MachineLearning

  23. The #Wasserstein distance (#EMD), sliced Wasserstein distance (#SWD), and the #L2norm are common #metrics used to quantify the ‘distance’ between two distributions. This tutorial compares these three metrics and discusses their advantages and disadvantages.

    🌎 fabriziomusacchio.com/blog/202

    #OptimalTransport #MachineLearning

  24. The #Wasserstein distance (#EMD), sliced Wasserstein distance (#SWD), and the #L2norm are common #metrics used to quantify the ‘distance’ between two distributions. This tutorial compares these three metrics and discusses their advantages and disadvantages.

    🌎 fabriziomusacchio.com/blog/202

    #OptimalTransport #MachineLearning

  25. This tutorial takes a different approach to explain the #Wasserstein distance (#EMD) by approximating the #EMD with cumulative distribution functions (#CDF), providing a more intuitive understanding of the metric.

    🌎 fabriziomusacchio.com/blog/202

    #OptimalTransport

  26. This tutorial takes a different approach to explain the #Wasserstein distance (#EMD) by approximating the #EMD with cumulative distribution functions (#CDF), providing a more intuitive understanding of the metric.

    🌎 fabriziomusacchio.com/blog/202

    #OptimalTransport

  27. This tutorial takes a different approach to explain the #Wasserstein distance (#EMD) by approximating the #EMD with cumulative distribution functions (#CDF), providing a more intuitive understanding of the metric.

    🌎 fabriziomusacchio.com/blog/202

    #OptimalTransport

  28. This tutorial takes a different approach to explain the #Wasserstein distance (#EMD) by approximating the #EMD with cumulative distribution functions (#CDF), providing a more intuitive understanding of the metric.

    🌎 fabriziomusacchio.com/blog/202

    #OptimalTransport

  29. Calculating the #Wasserstein distance (#EMD) 📈 can be computational costly when using #LinearProgramming. The #Sinkhorn algorithm provides a computationally efficient method for approximating the EMD, making it a practical choice for many applications, especially for large datasets 💫. Here is another tutorial, showing how to solve #OptimalTransport problem using the Sinkhorn algorithm in #Python 🐍

    🌎 fabriziomusacchio.com/blog/202

  30. Calculating the #Wasserstein distance (#EMD) 📈 can be computational costly when using #LinearProgramming. The #Sinkhorn algorithm provides a computationally efficient method for approximating the EMD, making it a practical choice for many applications, especially for large datasets 💫. Here is another tutorial, showing how to solve #OptimalTransport problem using the Sinkhorn algorithm in #Python 🐍

    🌎 fabriziomusacchio.com/blog/202

  31. Calculating the #Wasserstein distance (#EMD) 📈 can be computational costly when using #LinearProgramming. The #Sinkhorn algorithm provides a computationally efficient method for approximating the EMD, making it a practical choice for many applications, especially for large datasets 💫. Here is another tutorial, showing how to solve #OptimalTransport problem using the Sinkhorn algorithm in #Python 🐍

    🌎 fabriziomusacchio.com/blog/202

  32. Calculating the #Wasserstein distance (#EMD) 📈 can be computational costly when using #LinearProgramming. The #Sinkhorn algorithm provides a computationally efficient method for approximating the EMD, making it a practical choice for many applications, especially for large datasets 💫. Here is another tutorial, showing how to solve #OptimalTransport problem using the Sinkhorn algorithm in #Python 🐍

    🌎 fabriziomusacchio.com/blog/202

  33. The #Wasserstein distance 📐, aka Earth Mover’s Distance (#EMD), provides a robust and insightful approach for comparing #ProbabilityDistributions 📊. I’ve composed a #Python tutorial 🐍 that explains the #OptimalTransport problem required to calculate EMD. It also shows how to solve the OT problem and calculate the EMD using the Python Optimal Transport (POT) library. Feel free to use and share it 🤗

    🌎 fabriziomusacchio.com/blog/202

  34. The #Wasserstein distance 📐, aka Earth Mover’s Distance (#EMD), provides a robust and insightful approach for comparing #ProbabilityDistributions 📊. I’ve composed a #Python tutorial 🐍 that explains the #OptimalTransport problem required to calculate EMD. It also shows how to solve the OT problem and calculate the EMD using the Python Optimal Transport (POT) library. Feel free to use and share it 🤗

    🌎 fabriziomusacchio.com/blog/202

  35. The #Wasserstein distance 📐, aka Earth Mover’s Distance (#EMD), provides a robust and insightful approach for comparing #ProbabilityDistributions 📊. I’ve composed a #Python tutorial 🐍 that explains the #OptimalTransport problem required to calculate EMD. It also shows how to solve the OT problem and calculate the EMD using the Python Optimal Transport (POT) library. Feel free to use and share it 🤗

    🌎 fabriziomusacchio.com/blog/202

  36. The #Wasserstein distance 📐, aka Earth Mover’s Distance (#EMD), provides a robust and insightful approach for comparing #ProbabilityDistributions 📊. I’ve composed a #Python tutorial 🐍 that explains the #OptimalTransport problem required to calculate EMD. It also shows how to solve the OT problem and calculate the EMD using the Python Optimal Transport (POT) library. Feel free to use and share it 🤗

    🌎 fabriziomusacchio.com/blog/202

  37. #OptimalTransport: Moving stuff through a #labyrinth

    (Nicolas Papadakis: Optimal Transport for Image Processing, Signal and Image Processing. Université
    de Bordeaux; Habilitation thesis, 2015. tel-01246096v8, 2007)

  38. #OptimalTransport: Moving stuff through a #labyrinth

    (Nicolas Papadakis: Optimal Transport for Image Processing, Signal and Image Processing. Université
    de Bordeaux; Habilitation thesis, 2015. tel-01246096v8, 2007)

  39. #OptimalTransport: Moving stuff through a #labyrinth

    (Nicolas Papadakis: Optimal Transport for Image Processing, Signal and Image Processing. Université
    de Bordeaux; Habilitation thesis, 2015. tel-01246096v8, 2007)

  40. #OptimalTransport: Moving stuff through a #labyrinth

    (Nicolas Papadakis: Optimal Transport for Image Processing, Signal and Image Processing. Université
    de Bordeaux; Habilitation thesis, 2015. tel-01246096v8, 2007)

  41. Our Pick of the week: Phuong-Hang Le et al., "Pre-training for Speech Translation: CTC Meets Optimal Transport"
    by @mgaido91

    :arxiv: arxiv.org/abs/2301.11716

    #NLProc #optimaltransport #CTC #speechtranslation

  42. Our Pick of the week: Phuong-Hang Le et al., "Pre-training for Speech Translation: CTC Meets Optimal Transport"
    by @mgaido91

    :arxiv: arxiv.org/abs/2301.11716

    #NLProc #optimaltransport #CTC #speechtranslation

  43. Our Pick of the week: Phuong-Hang Le et al., "Pre-training for Speech Translation: CTC Meets Optimal Transport"
    by @mgaido91

    :arxiv: arxiv.org/abs/2301.11716

    #NLProc #optimaltransport #CTC #speechtranslation

  44. Our Pick of the week: Phuong-Hang Le et al., "Pre-training for Speech Translation: CTC Meets Optimal Transport"
    by @mgaido91

    :arxiv: arxiv.org/abs/2301.11716

    #NLProc #optimaltransport #CTC #speechtranslation

  45. The first is Éric Daoud, who defended on Monday.

    His dissertation is titled "Geographic and socio-demographic disparities in oncology care pathways" and was supervised by Fabien Reyal and Marc Lelarge. Éric was a member of a working group we had on #machineLearning and #electronicHealthRecords until I was sick. See his preprints here: edaoud.com/research/ Interesting applications of #optimalTransport inside!

    2/3

  46. The first is Éric Daoud, who defended on Monday.

    His dissertation is titled "Geographic and socio-demographic disparities in oncology care pathways" and was supervised by Fabien Reyal and Marc Lelarge. Éric was a member of a working group we had on #machineLearning and #electronicHealthRecords until I was sick. See his preprints here: edaoud.com/research/ Interesting applications of #optimalTransport inside!

    2/3

  47. The first is Éric Daoud, who defended on Monday.

    His dissertation is titled "Geographic and socio-demographic disparities in oncology care pathways" and was supervised by Fabien Reyal and Marc Lelarge. Éric was a member of a working group we had on #machineLearning and #electronicHealthRecords until I was sick. See his preprints here: edaoud.com/research/ Interesting applications of #optimalTransport inside!

    2/3

  48. The first is Éric Daoud, who defended on Monday.

    His dissertation is titled "Geographic and socio-demographic disparities in oncology care pathways" and was supervised by Fabien Reyal and Marc Lelarge. Éric was a member of a working group we had on #machineLearning and #electronicHealthRecords until I was sick. See his preprints here: edaoud.com/research/ Interesting applications of #optimalTransport inside!

    2/3

  49. The first is Éric Daoud, who defended on Monday.

    His dissertation is titled "Geographic and socio-demographic disparities in oncology care pathways" and was supervised by Fabien Reyal and Marc Lelarge. Éric was a member of a working group we had on #machineLearning and #electronicHealthRecords until I was sick. See his preprints here: edaoud.com/research/ Interesting applications of #optimalTransport inside!

    2/3