際際滷shows by User: PengCheng2 / http://www.slideshare.net/images/logo.gif 際際滷shows by User: PengCheng2 / Sun, 20 Feb 2022 23:25:06 GMT 際際滷Share feed for 際際滷shows by User: PengCheng2 Shape Safety in Tensor Programming is Easy for a Theorem Prover -SBTB 2021 /slideshow/shape-safety-in-tensor-programming-is-easy-for-a-theorem-prover-sbtb-2021/251210680 slide-220220232506
We present shapesafe (https://github.com/tribbloid/shapesafe) - the most comprehensive compile-time verifier for scala linear algebra - by only exploiting scala's type system as a theorem prover. This new paradigm allows type-level tensor computations, even those as complex as composite neural network blocks, to be rewritten, simplified and verified while being written. We will talk about its design and limitations, and most important, what we have observed and learned from it]]>
We present shapesafe (https://github.com/tribbloid/shapesafe) - the most comprehensive compile-time verifier for scala linear algebra - by only exploiting scala's type system as a theorem prover. This new paradigm allows type-level tensor computations, even those as complex as composite neural network blocks, to be rewritten, simplified and verified while being written. We will talk about its design and limitations, and most important, what we have observed and learned from it]]> Sun, 20 Feb 2022 23:25:06 GMT /slideshow/shape-safety-in-tensor-programming-is-easy-for-a-theorem-prover-sbtb-2021/251210680 PengCheng2@slideshare.net(PengCheng2) Shape Safety in Tensor Programming is Easy for a Theorem Prover -SBTB 2021 PengCheng2 We present shapesafe (https://github.com/tribbloid/shapesafe) - the most comprehensive compile-time verifier for scala linear algebra - by only exploiting scala's type system as a theorem prover. This new paradigm allows type-level tensor computations, even those as complex as composite neural network blocks, to be rewritten, simplified and verified while being written. We will talk about its design and limitations, and most important, what we have observed and learned from it <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/slide-220220232506-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> We present shapesafe (https://github.com/tribbloid/shapesafe) - the most comprehensive compile-time verifier for scala linear algebra - by only exploiting scala&#39;s type system as a theorem prover. This new paradigm allows type-level tensor computations, even those as complex as composite neural network blocks, to be rewritten, simplified and verified while being written. We will talk about its design and limitations, and most important, what we have observed and learned from it
Shape Safety in Tensor Programming is Easy for a Theorem Prover -SBTB 2021 from Peng Cheng
]]> 112 0 https://cdn.slidesharecdn.com/ss_thumbnails/slide-220220232506-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 Tensor Field Network (and other ConvNet Generalisations) /slideshow/tensor-field-network-and-other-convnet-generalisations-132342506/132342506 out-190219021322
In this session we will discuss the relationship between data augmentation, invariant/equivariant features and the abstract concept of convolution layers, specifically, how this abstraction can be extended to devise concrete neural network architectures that are robust to diverse data and augmentation types (all of which are published after 2016). We will focus on the latest of the series, designed to handle spatial graph data augmentable by 3d translations & rotations. In the end, we will showcase it's applications in molecule analysis and autonomous flight. We will focus on the latest of the series, designed to handle spatial graph data augmentable by 3d translations & rotations. In the end, we will showcase it's applications in molecule analysis and autonomous flight.]]>
In this session we will discuss the relationship between data augmentation, invariant/equivariant features and the abstract concept of convolution layers, specifically, how this abstraction can be extended to devise concrete neural network architectures that are robust to diverse data and augmentation types (all of which are published after 2016). We will focus on the latest of the series, designed to handle spatial graph data augmentable by 3d translations & rotations. In the end, we will showcase it's applications in molecule analysis and autonomous flight. We will focus on the latest of the series, designed to handle spatial graph data augmentable by 3d translations & rotations. In the end, we will showcase it's applications in molecule analysis and autonomous flight.]]> Tue, 19 Feb 2019 02:13:22 GMT /slideshow/tensor-field-network-and-other-convnet-generalisations-132342506/132342506 PengCheng2@slideshare.net(PengCheng2) Tensor Field Network (and other ConvNet Generalisations) PengCheng2 In this session we will discuss the relationship between data augmentation, invariant/equivariant features and the abstract concept of convolution layers, specifically, how this abstraction can be extended to devise concrete neural network architectures that are robust to diverse data and augmentation types (all of which are published after 2016). We will focus on the latest of the series, designed to handle spatial graph data augmentable by 3d translations & rotations. In the end, we will showcase it's applications in molecule analysis and autonomous flight. We will focus on the latest of the series, designed to handle spatial graph data augmentable by 3d translations & rotations. In the end, we will showcase it's applications in molecule analysis and autonomous flight. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/out-190219021322-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> In this session we will discuss the relationship between data augmentation, invariant/equivariant features and the abstract concept of convolution layers, specifically, how this abstraction can be extended to devise concrete neural network architectures that are robust to diverse data and augmentation types (all of which are published after 2016). We will focus on the latest of the series, designed to handle spatial graph data augmentable by 3d translations &amp; rotations. In the end, we will showcase it&#39;s applications in molecule analysis and autonomous flight. We will focus on the latest of the series, designed to handle spatial graph data augmentable by 3d translations &amp; rotations. In the end, we will showcase it&#39;s applications in molecule analysis and autonomous flight.
Tensor Field Network (and other ConvNet Generalisations) from Peng Cheng
]]> 352 1 https://cdn.slidesharecdn.com/ss_thumbnails/out-190219021322-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 A fast graphic api for non-linear machine learning /slideshow/a-fast-graphic-api-for-nonlinear-machine-learning/61780523 afastgraphicapifornon-linearmachinelearning-160507234553
Draft v1]]>
Draft v1]]> Sat, 07 May 2016 23:45:52 GMT /slideshow/a-fast-graphic-api-for-nonlinear-machine-learning/61780523 PengCheng2@slideshare.net(PengCheng2) A fast graphic api for non-linear machine learning PengCheng2 Draft v1 <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/afastgraphicapifornon-linearmachinelearning-160507234553-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Draft v1
A fast graphic api for non-linear machine learning from Peng Cheng
]]> 574 7 https://cdn.slidesharecdn.com/ss_thumbnails/afastgraphicapifornon-linearmachinelearning-160507234553-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 Query breakdown /slideshow/query-breakdown/38464291 querybreakdown-140828141443-phpapp01
A visual breakdown of Query in Spookystuff, a data scraping/mashup engine]]>
A visual breakdown of Query in Spookystuff, a data scraping/mashup engine]]> Thu, 28 Aug 2014 14:14:43 GMT /slideshow/query-breakdown/38464291 PengCheng2@slideshare.net(PengCheng2) Query breakdown PengCheng2 A visual breakdown of Query in Spookystuff, a data scraping/mashup engine <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/querybreakdown-140828141443-phpapp01-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> A visual breakdown of Query in Spookystuff, a data scraping/mashup engine
Query breakdown from Peng Cheng
]]> 386 7 https://cdn.slidesharecdn.com/ss_thumbnails/querybreakdown-140828141443-phpapp01-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 How to build your query engine in spark /slideshow/how-to-build-your-query-engine-in-spark/37448577 howtobuildyourqueryengineinspark-140729020034-phpapp01
An over-ambitious introduction to Spark programming, test and deployment. This slide tries to cover most core technologies and design patterns used in SpookyStuff, the fastest query engine for data collection/mashup from the deep web. For more information please follow: https://github.com/tribbloid/spookystuff A bug in PowerPoint used to cause transparent background color not being rendered properly. This has been fixed in a recent upload.]]>
An over-ambitious introduction to Spark programming, test and deployment. This slide tries to cover most core technologies and design patterns used in SpookyStuff, the fastest query engine for data collection/mashup from the deep web. For more information please follow: https://github.com/tribbloid/spookystuff A bug in PowerPoint used to cause transparent background color not being rendered properly. This has been fixed in a recent upload.]]> Tue, 29 Jul 2014 02:00:33 GMT /slideshow/how-to-build-your-query-engine-in-spark/37448577 PengCheng2@slideshare.net(PengCheng2) How to build your query engine in spark PengCheng2 An over-ambitious introduction to Spark programming, test and deployment. This slide tries to cover most core technologies and design patterns used in SpookyStuff, the fastest query engine for data collection/mashup from the deep web. For more information please follow: https://github.com/tribbloid/spookystuff A bug in PowerPoint used to cause transparent background color not being rendered properly. This has been fixed in a recent upload. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/howtobuildyourqueryengineinspark-140729020034-phpapp01-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> An over-ambitious introduction to Spark programming, test and deployment. This slide tries to cover most core technologies and design patterns used in SpookyStuff, the fastest query engine for data collection/mashup from the deep web. For more information please follow: https://github.com/tribbloid/spookystuff A bug in PowerPoint used to cause transparent background color not being rendered properly. This has been fixed in a recent upload.
How to build your query engine in spark from Peng Cheng
]]> 4698 5 https://cdn.slidesharecdn.com/ss_thumbnails/howtobuildyourqueryengineinspark-140729020034-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 https://cdn.slidesharecdn.com/profile-photo-PengCheng2-48x48.jpg?cb=1682453735 https://cdn.slidesharecdn.com/ss_thumbnails/slide-220220232506-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/shape-safety-in-tensor-programming-is-easy-for-a-theorem-prover-sbtb-2021/251210680 Shape Safety in Tensor... https://cdn.slidesharecdn.com/ss_thumbnails/out-190219021322-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/tensor-field-network-and-other-convnet-generalisations-132342506/132342506 Tensor Field Network (... https://cdn.slidesharecdn.com/ss_thumbnails/afastgraphicapifornon-linearmachinelearning-160507234553-thumbnail.jpg?width=320&height=320&fit=bounds slideshow/a-fast-graphic-api-for-nonlinear-machine-learning/61780523 A fast graphic api for...