]> 2020-04-29 A Comprehensive Survey of Knowledge Graph Embeddings with Literals: Techniques and Applications > survey of the KG embedding models which consider the structured information of the graph as well as the unstructured information in form of literals such as text, numerical values etc A le mérite de poser la question de l'utilisation de littéraux dans les embeddings de KG [Newer and longer version](/doc/2020/05/1910_12507_a_survey_on_knowle) CEUR-WS.org/Vol-2377 - Workshop on Deep Learning for Knowledge Graphs 2019 2020-04-29T14:04:51Z 2020-04-29 2020-04-29T14:09:42Z Coronavirus en France : « En matière de prévention, nous ne sommes pas à la hauteur de l’épidémie » 2020-04-11T14:45:02Z 2020-04-11 > Arguing that you don't care about the right to privacy because you have nothing to hide is no different than saying you don't care about free speech because you have nothing to say. > When you say, 'I have nothing to hide,' you're saying, 'I don't care about this right'. E. Snowden Nothing to hide argument 2020-04-24 2020-04-24T12:39:27Z 2020-04-27T19:39:48Z 2020-04-27 Andrej Karpathy | Multi-Task Learning in the Wilderness · SlidesLive Dans les quartiers populaires, « si on remplit le frigo, on chope le corona » 2020-04-18 2020-04-18T16:00:15Z > Les quartiers populaires entament leur deuxième mois de confinement à bout de souffle 2020-04-22 Nicolas Pinto Jesse Thomason Successful linguistic communication relies on a shared experience of the world, and it is this shared experience that makes utterances meaningful. Despite the incredible effectiveness of language processing models trained on text alone, today's best systems still make mistakes that arise from a failure to relate language to the physical world it describes and to the social interactions it facilitates. Natural Language Processing is a diverse field, and progress throughout its development has come from new representational theories, modeling techniques, data collection paradigms, and tasks. We posit that the present success of representation learning approaches trained on large text corpora can be deeply enriched from the parallel tradition of research on the contextual and social nature of language. In this article, we consider work on the contextual foundations of language: grounding, embodiment, and social interaction. We describe a brief history and possible progression of how contextual information can factor into our representations, with an eye towards how this integration can move the field forward and where it is currently being pioneered. We believe this framing will serve as a roadmap for truly contextual language understanding. Joseph Turian Experience Grounds Language 2020-04-21T16:56:27Z Angeliki Lazaridou Jacob Andreas [2004.10151] Experience Grounds Language 2004.10151 Yonatan Bisk 2020-04-21T16:56:27Z Mirella Lapata Yonatan Bisk Yoshua Bengio Jonathan May 2020-04-22T16:52:37Z Aleksandr Nisnevich Joyce Chai Ari Holtzman 2020-04-25T21:35:21Z 2020-04-25 « La gestion de la pandémie de Covid-19 et les mesures nécessaires à la sortie de crise conspirent à faire de l’environnement une question subsidiaire » Pour éradiquer la lucilie bouchère en Libye, des mouches mâles stériles sont lâchées vers les femelles dont les larves dévorent le bétail (1991) 2020-04-03 2020-04-03T19:20:18Z Les URLs du Monde ne sont pas cool 2020-04-25T13:07:26Z 2020-04-25 from "(Ai-je le droit de mettre un lien vers Le Monde.fr sur mon site ?": > Attention : l’URL des articles est modifiée au moment de leur passage en archive payante. c'est lamentable, et je ne renouvellerai donc pas mon abonnement, parce que je ne peux pas me servir du Monde comme d'une base de connaissance. "Cool URIs don't change" ([#TBL](/tag/tim_berners_lee)) 2020-04-18T13:33:29Z 2020-04-18 COVID-19 Lessons from Three Mile Island #2 — the NRC | I, Cringely Knowledge Distillation - Neural Network Distiller (Distiller is an open-source Python package for neural network compression research. The doc about knowledge distillation) 2020-04-22T21:52:26Z 2020-04-22 pdf2table: A Method to Extract Table Information from PDF Files 2020-04-02 2020-04-02T15:35:47Z 2020-04-14T23:49:35Z 2020-04-14 How do different communities create unique identifiers? – Lost Boy so old a question Coronavirus: 38 days when Britain sleepwalked into disaster | News | The Sunday Times 2020-04-19 2020-04-19T12:20:40Z 2020-04-04T10:57:29Z 2020-04-04 En pleine crise sanitaire, le géant américain Palantir lorgne les données des hôpitaux français « Les cas de Covid se multiplient. Ça tombe, ça tombe. Jusqu’où ? » : la course à la vie d’une réanimatrice 2020-04-06 2020-04-06T14:29:57Z 2020-04-08T21:08:07Z 2020-04-08 Curiosités animales - Une vie sans sexe : le dragon de Komodo et le puceron | ARTE BrunoRB/ahocorasick: Aho-corasick for javascript. 2020-04-18 2020-04-18T00:37:31Z Arundhati Roy : « En Inde, le confinement le plus gigantesque et le plus punitif de la planète » 2020-04-06 2020-04-06T19:31:38Z Copier le fonctionnement du cerveau pour économiser de l'énergie | Techniques de l'Ingénieur 2020-04-05 2020-04-05T11:14:51Z Projota - A Rezadeira (Video Oficial) - YouTube 2020-04-10 2020-04-10T19:30:37Z 2020-04-16 2020-04-16T19:15:38Z How many words—and which ones—are sufficient to define all other words? The Latent Structure of Dictionaries - Vincent‐Lamarre - 2016 2020-04-04T14:39:01Z 2020-04-04 Coronavirus : la Seine-Saint-Denis confrontée à une inquiétante surmortalité « Les infirmières, les caissières, les aides-soignantes, les agents d’entretien, les intérimaires, les agents de sécurité, les livreurs… bref, **tous ceux qui font tenir la France debout aujourd’hui, tous ceux qui vont au front et se mettent en danger**, ils viennent des quartiers populaires, ce sont des habitants du 93 ! » 2020-04-10T17:54:09Z [2004.05150] Longformer: The Long-Document Transformer Matthew E. Peters Transformer-based models are unable to process long sequences due to their self-attention operation, which scales quadratically with the sequence length. To address this limitation, we introduce the Longformer with an attention mechanism that scales linearly with sequence length, making it easy to process documents of thousands of tokens or longer. Longformer's attention mechanism is a drop-in replacement for the standard self-attention and combines a local windowed attention with a task motivated global attention. Following prior work on long-sequence transformers, we evaluate Longformer on character-level language modeling and achieve state-of-the-art results on text8 and enwik8. In contrast to most prior work, we also pretrain Longformer and finetune it on a variety of downstream tasks. Our pretrained Longformer consistently outperforms RoBERTa on long document tasks and sets new state-of-the-art results on WikiHop and TriviaQA. Arman Cohan > **Transformer-based models are unable to process long sequences due to their self-attention operation, which scales quadratically with the sequence length**. To address this limitation, we introduce the Longformer with an attention mechanism that scales linearly with sequence length. Longformer's attention mechanism is a **drop-in replacement** for the standard self-attention and **combines a local windowed attention with a task motivated global attention**. Iz Beltagy 2004.05150 2020-04-13 Longformer: The Long-Document Transformer 2020-04-10T17:54:09Z 2020-04-13T11:06:40Z Iz Beltagy Julian Eisenschlos sur Twitter : "Ever wondered how to pre-train models that understand tables and do QA?" 2020-04-11 2020-04-11T17:55:37Z Contrastive Predictive Coding 2020-04-25T17:40:06Z 2020-04-25 Keras implementation of the algorithm presented in this DeepMind's [paper](/doc/?uri=https%3A%2F%2Farxiv.org%2Fabs%2F1807.03748) Jiaming Shen [2001.09522] TaxoExpan: Self-supervised Taxonomy Expansion with Position-Enhanced Graph Neural Network Zhihong Shen 2020-01-26T21:30:21Z Kuansan Wang 2020-01-26T21:30:21Z Jiaming Shen Chenyan Xiong 2020-04-25T10:03:35Z how to add a set of new concepts to an existing taxonomy. [Tweet](https://twitter.com/mickeyjs6/status/1253772146142216194?s=20) [GitHub](https://github.com/mickeystroller/TaxoExpan) > we study the taxonomy expansion task: given an existing taxonomy and a set of new emerging concepts, we aim to automatically expand the taxonomy to incorporate these new concepts (without changing the existing relations in the given taxonomy). > To the best of our knowledge, this is the first study on **how to expand an existing directed acyclic graph (as we model a taxonomy as a DAG) using self-supervised learning**. Self-supervised framework, the existing taxonomy being used as training data: it learns a model to predict whether a query concept is the direct hyponym of an anchor concept. > 2 techniques: > > 1. a **position-enhanced graph neural network that encodes the local structure of an anchor concept** in the existing taxonomy, > 2. a noise-robust training objective that enables the learned model to be insensitive to the label noise in the self-supervision data. Regarding 1: uses [GNN](/tag/graph_neural_networks.html) to model the "ego network" of concepts (potential “siblings” and “grand parents” of the query concept). > Regular GNNs fail to distinguish nodes with different relative positions to the query (i.e., some nodes are grand parents of the query while the others are siblings of the query). To address this limitation, we present a simple but effective enhancement to inject such position information into GNNs using position embedding. We show that such embedding can be easily integrated with existing GNN architectures (e.g., [GCN](/tag/graph_convolutional_networks) and GAT) and significantly boosts the prediction performance Regarding point 2: uses InfoNCE loss, cf. [Contrastive Predictive Coding](/doc/?uri=https%3A%2F%2Farxiv.org%2Fabs%2F1807.03748) > Instead of predicting whether each individual ⟨query concept, anchor concept⟩ pair is positive or not, we first group all pairs sharing the same query concept into a single training instance and learn a model to select the positive pair among other negative ones from the group. (Hum, ça me rappelle quelque chose) > assume each concept (in existing taxonomy + set of new concepts) has an initial embedding vector learned from some text associated with this concept. To keep things tractable, only attempts to find a single parent node of each new concept. 2020-04-25 Chi Wang Jiawei Han TaxoExpan: Self-supervised Taxonomy Expansion with Position-Enhanced Graph Neural Network 2001.09522 Taxonomies consist of machine-interpretable semantics and provide valuable knowledge for many web applications. For example, online retailers (e.g., Amazon and eBay) use taxonomies for product recommendation, and web search engines (e.g., Google and Bing) leverage taxonomies to enhance query understanding. Enormous efforts have been made on constructing taxonomies either manually or semi-automatically. However, with the fast-growing volume of web content, existing taxonomies will become outdated and fail to capture emerging knowledge. Therefore, in many applications, dynamic expansions of an existing taxonomy are in great demand. In this paper, we study how to expand an existing taxonomy by adding a set of new concepts. We propose a novel self-supervised framework, named TaxoExpan, which automatically generates a set of <query concept, anchor concept> pairs from the existing taxonomy as training data. Using such self-supervision data, TaxoExpan learns a model to predict whether a query concept is the direct hyponym of an anchor concept. We develop two innovative techniques in TaxoExpan: (1) a position-enhanced graph neural network that encodes the local structure of an anchor concept in the existing taxonomy, and (2) a noise-robust training objective that enables the learned model to be insensitive to the label noise in the self-supervision data. Extensive experiments on three large-scale datasets from different domains demonstrate both the effectiveness and the efficiency of TaxoExpan for taxonomy expansion. 2020-04-10T15:27:25Z 2020-04-10 Barbara Stiegler : « La crise due au coronavirus reflète la vision néolibérale de la santé publique » 1904.01947 2020-04-02T15:48:47Z 2019-04-03T12:12:03Z Extracting information from tables in documents presents a significant challenge in many industries and in academic research. Existing methods which take a bottom-up approach of integrating lines into cells and rows or columns neglect the available prior information relating to table structure. Our proposed method takes a top-down approach, first using a generative adversarial network to map a table image into a standardised `skeleton' table form denoting the approximate row and column borders without table content, then fitting renderings of candidate latent table structures to the skeleton structure using a distance measure optimised by a genetic algorithm. Mark Rowan [1904.01947] Extracting Tables from Documents using Conditional Generative Adversarial Networks and Genetic Algorithms 2020-04-02 Nataliya Le Vine 2019-04-03T12:12:03Z Matthew Zeigenfuse Nataliya Le Vine Extracting Tables from Documents using Conditional Generative Adversarial Networks and Genetic Algorithms 2020-04-11T14:47:15Z 2020-04-11 Les leçons de Wuhan pour enrayer l’épidémie 2020-04-15T00:50:04Z 2020-04-15 Finally We May Have a Path to the Fundamental Theory of Physics… and It’s Beautiful—Stephen Wolfram Writings Climat : le patronat s’active pour infléchir les normes 2020-04-25 2020-04-25T21:38:05Z TL;DR: yes 2020-04-27T15:33:37Z 2020-04-27 Should you use FastAI? - deeplearningbrasilia - Medium 2020-04-02 2020-04-02T15:38:47Z pdftables.com: PDF to Excel converter - PDFTables Retail Graph — Walmart’s Product Knowledge Graph 2020-04-09 2020-04-09T21:18:53Z 2020-04-29 Iterative Entity Alignment with Improved Neural Attribute Embedding 2020-04-29T19:04:03Z 2020-04-04T11:02:50Z 2020-04-04 Damien Henry sur Twitter : "This code is so beautifully written, it almost hurts." 1903.04197 2020-04-16T14:13:03Z Changyong Shun 2020-04-16 2019-03-11T10:05:09Z Chunhua Shen 2020-02-20T23:52:50Z Structured Knowledge Distillation for Dense Prediction [1903.04197] Structured Knowledge Distillation for Dense Prediction Yifan Liu Jingdong Wang Yifan Liu In this paper, we consider transferring the structure information from large networks to small ones for dense prediction tasks. Previous knowledge distillation strategies used for dense prediction tasks often directly borrow the distillation scheme for image classification and perform knowledge distillation for each pixel separately, leading to sub-optimal performance. Here we propose to distill structured knowledge from large networks to small networks, taking into account the fact that dense prediction is a structured prediction problem. Specifically, we study two structured distillation schemes: i)pair-wise distillation that distills the pairwise similarities by building a static graph, and ii)holistic distillation that uses adversarial training to distill holistic knowledge. The effectiveness of our knowledge distillation approaches is demonstrated by extensive experiments on three dense prediction tasks: semantic segmentation, depth estimation, and object detection. 2020-04-24 Dépistage du coronavirus : les raisons du fiasco français sur les tests 2020-04-24T14:00:52Z [doc](https://rdflib.readthedocs.io/en/stable/) RDFLib/rdflib: a Python library for working with RDF 2020-04-09 2020-04-09T01:56:54Z 7 Alternatives to the div HTML Tag - Zac Heisey - Medium 2020-04-27 2020-04-27T15:34:28Z > les carences dans le mode de pensée, jointes à la domination incontestable d’une soif effrénée de profit, sont responsables d’innombrables désastres humains dont ceux survenus depuis février 2020 2020-04-19T15:50:59Z 2020-04-19 Edgar Morin : « Cette crise nous pousse à nous interroger sur notre mode de vie, sur nos vrais besoins masqués dans les aliénations du quotidien » > TLDR - it's excel for text + a graph database for your ideas Roam Research – A note taking tool for networked thought. 2020-04-17 2020-04-17T13:54:47Z Lee Moses - Bad Girl (full song, no break) - YouTube 2020-04-13 2020-04-13T15:02:28Z Blog de Raphaël Sourty 2020-04-29 2020-04-29T16:43:58Z 2020-04-12T11:38:36Z 2020-04-12 Lainey Doyle sur Twitter : Basic things: Ireland and the UK started this pandemic with roughly the same..." Kin Sum Liu Chien-Chun Ni 2004.06842 Chien-Chun Ni 2020-04-17T19:14:01Z 2020-04-15T00:49:27Z In this paper, we describe an embedding-based entity recommendation framework for Wikipedia that organizes Wikipedia into a collection of graphs layered on top of each other, learns complementary entity representations from their topology and content, and combines them with a lightweight learning-to-rank approach to recommend related entities on Wikipedia. Through offline and online evaluations, we show that the resulting embeddings and recommendations perform well in terms of quality and user engagement. Balancing simplicity and quality, this framework provides default entity recommendations for English and other languages in the Yahoo! Knowledge Graph, which Wikipedia is a core subset of. Layered Graph Embedding for Entity Recommendation using Wikipedia in the Yahoo! Knowledge Graph [2004.06842] Layered Graph Embedding for Entity Recommendation using Wikipedia in the Yahoo! Knowledge Graph 2020-04-15T00:49:27Z Nicolas Torzec an embedding-based entity recommendation framework for Wikipedia that organizes Wikipedia into a collection of graphs layered on top of each other, **learns complementary entity representations from their topology and content**, and combines them with a lightweight **learning-to-rank** approach to recommend related entities on Wikipedia 2020-04-17 Turning Up the Heat: The Mechanics of Model Distillation 2020-04-22T21:40:13Z 1503.02531 [1503.02531] Distilling the Knowledge in a Neural Network Geoffrey Hinton 2020-04-16 2020-04-16T14:40:33Z A very simple way to improve the performance of almost any machine learning algorithm is to train many different models on the same data and then to average their predictions. Unfortunately, making predictions using a whole ensemble of models is cumbersome and may be too computationally expensive to allow deployment to a large number of users, especially if the individual models are large neural nets. Caruana and his collaborators have shown that it is possible to compress the knowledge in an ensemble into a single model which is much easier to deploy and we develop this approach further using a different compression technique. We achieve some surprising results on MNIST and we show that we can significantly improve the acoustic model of a heavily used commercial system by distilling the knowledge in an ensemble of models into a single model. We also introduce a new type of ensemble composed of one or more full models and many specialist models which learn to distinguish fine-grained classes that the full models confuse. Unlike a mixture of experts, these specialist models can be trained rapidly and in parallel. Distilling the Knowledge in a Neural Network 2015-03-09T15:44:49Z Geoffrey Hinton Jeff Dean 2015-03-09T15:44:49Z > **a different kind of training**, which we call “**distillation**” to transfer the knowledge from the cumbersome model to a small model that is more suitable for deployment > Caruana and his collaborators have shown that it is possible to compress the knowledge in an [#ensemble](/tag/ensemble_learning.html) into a single model which is much easier to deploy and we develop this approach further using a different compression technique. We achieve some surprising results on MNIST. Oriol Vinyals About the [2015 Hinton's paper](/doc/2020/04/1503_02531_distilling_the_kno) > The role of temperature is to push the model into a region where it’s emitting less extreme probabilities, so that they’re more informative to our calculations 2020-04-22 Knowledge Graphs @ ICLR 2020 - Michael Galkin - Medium 2020-04-28 2020-04-28T08:29:22Z 1. Neural Reasoning for Complex QA with KGs 2. KG-augmented Language Models 3. KG Embeddings: Temporal and Inductive Inference 4. Entity Matching with GNNs 2020-04-14T21:22:47Z 2020-04-14 Camel Express News April 2020 1995-11-29T19:32:04Z 2020-04-27T17:22:44Z Philip Resnik 1995-11-29T19:32:04Z Philip Resnik 2020-04-27 This paper presents a new measure of semantic similarity in an IS-A taxonomy, based on the notion of information content. Experimental evaluation suggests that the measure performs encouragingly well (a correlation of r = 0.79 with a benchmark set of human similarity judgments, with an upper bound of r = 0.90 for human subjects performing the same task), and significantly better than the traditional edge counting approach (r = 0.66). cmp-lg/9511007 Using Information Content to Evaluate Semantic Similarity in a Taxonomy [cmp-lg/9511007] Using Information Content to Evaluate Semantic Similarity in a Taxonomy (1995) Learning Attention-based Embeddings for Relation Prediction in Knowledge Graphs 1906.01195 2019-06-04T04:59:08Z Deepak Nathani 2020-04-30 Jatin Chauhan Manohar Kaul 2020-04-30T12:59:24Z Deepak Nathani The recent proliferation of knowledge graphs (KGs) coupled with incomplete or partial information, in the form of missing relations (links) between entities, has fueled a lot of research on knowledge base completion (also known as relation prediction). Several recent works suggest that convolutional neural network (CNN) based models generate richer and more expressive feature embeddings and hence also perform well on relation prediction. However, we observe that these KG embeddings treat triples independently and thus fail to cover the complex and hidden information that is inherently implicit in the local neighborhood surrounding a triple. To this effect, our paper proposes a novel attention based feature embedding that captures both entity and relation features in any given entity's neighborhood. Additionally, we also encapsulate relation clusters and multihop relations in our model. Our empirical study offers insights into the efficacy of our attention based model and we show marked performance gains in comparison to state of the art methods on all datasets. [1906.01195] Learning Attention-based Embeddings for Relation Prediction in Knowledge Graphs 2019-06-04T04:59:08Z Charu Sharma [GitHub](https://github.com/deepakn97/relationPrediction) [Blog post](/doc/2020/04/deepak_nathani_%7C_pay_attention_) Blog post for this [paper](/doc/2020/04/1906_01195_learning_attention) 2020-04-30T13:03:33Z 2020-04-30 Deepak Nathani | Pay Attention, Relations are Important