[{"data":1,"prerenderedAt":175},["ShallowReactive",2],{"project-12619694":3},{"id":4,"name":5,"line":6,"line_source":7,"title":8,"url":9,"domain":10,"maker":11,"posted_at":12,"year":13,"points":14,"comments":15,"rank_in_year":16,"shown_in_year":17,"top":18,"tier":19,"category":20,"subcategory":21,"format":22,"keywords":23,"reception":28,"sentiment":29,"stance_mix":30,"scored_comments":34,"aspects":35,"strip":38,"post_text":41,"what":42,"said":45,"maker_in_thread":100,"shape":109,"grid":112,"attempts":137,"same_week":138,"maker_shows":117,"maker_hidden":167,"reader_version":168,"card":169},"12619694","Graph Convolutional Networks","Intro to neural networks on graphs","title","Show HN: Graph Convolutional Networks – Intro to neural networks on graphs","http:\u002F\u002Ftkipf.github.io\u002Fgraph-convolutional-networks\u002F","tkipf.github.io","triplefloat",1475352981,2016,106,12,346,2264,16,"notable","ai","graph neural networks","library",[24,25,26,27],"graph convolutional networks","neural networks on graphs","graph neural network","node classification","curious",0.178,{"positive":31,"neutral":32,"mixed":33,"critical":33},2,4,0,6,{"idea":36,"execution":37},{"positive":31,"critical":33},{"positive":33,"critical":33},[39,33,40,33,40,33,40,39,40,33,33,33],5,3,null,{"problem":41,"audience":41,"pricing":41,"motivation":41,"time_to_build":41,"stack":43,"openness":44,"stage":44,"team_size":44},[],"unknown",{"praise":46,"objections":59,"requests":64,"questions":72,"alternatives":94},[47],{"value":48,"source":49,"evidence":50},"Powerful and interesting approach to learning on graphs","commenter",[51,55],{"id":52,"by":53,"text":54},"12621571","kingennio","This is really powerful. As a use case I'm considering the graph of streets in a city. A task would be to make predictions on the next nodes visited by a vehicle given the…",{"id":56,"by":57,"text":58},"12620787","empath75","This looks really interesting, but I'm struggling to think of a practical application. Any ideas?",[60],{"value":61,"source":49,"evidence":62},"Practical applications are hard to see",[63],{"id":56,"by":57,"text":58},[65],{"value":66,"source":49,"evidence":67},"Mini-batching support for large graphs",[68,69],{"id":52,"by":53,"text":54},{"id":70,"by":11,"text":71},"12621612","Mini-batching is indeed tricky, as you need the information of the complete local neighborhood for all nodes in a mini-batch. Let's say you select N nodes for a mini-batch, then…",[73,80,87],{"value":74,"source":49,"evidence":75},"Does it classify whole graphs or individual nodes",[76],{"id":77,"by":78,"text":79},"12620143","SubiculumCode","Does this classify whole graphs, or nodes into communities\u002Fclasses?",{"value":81,"source":49,"evidence":82},"How does it differ from graphical models such as factor graphs",[83],{"id":84,"by":85,"text":86},"12621101","webyqop","How does this differ from graphical models like factor graphs ?",{"value":88,"source":49,"evidence":89},"Can Bayesian networks be used as input",[90],{"id":91,"by":92,"text":93},"12621540","singularity2001","Stupid question: can bayesian networks (graphs) be a useful input?",[95],{"value":96,"source":49,"evidence":97},"Graphical models such as factor graphs and Bayesian networks",[98,99],{"id":84,"by":85,"text":86},{"id":91,"by":92,"text":93},{"comments":32,"answered":32,"key":101},[102,103,106],{"id":70,"text":71},{"id":104,"text":105},"12620172","Both graph-level and node-level classification are possible. Graph-level classification requires some from of pooling operation (simplest case: mean-pooling over all nodes, but…",{"id":107,"text":108},"12621161","In a graphical model, you'd explicitly model the probabilistic assumptions that you make with respect to the data. In this neural network-based approach the goal can be thought…",{"commenters":34,"max_depth":39,"maker_replies":32,"first_comment_mins":110,"lifespan_hours":111},126,74,[113,115,118,120,121,123,125,126,128,129,131,134],[52,33,39,114,53,33],"project\u002Fpositive\u002Fidea",[70,33,33,116,11,117],"maker\u002Fneutral\u002Fexecution",1,[77,117,40,119,78,33],"project\u002Fneutral\u002Fexecution",[104,117,33,116,11,117],[84,31,40,122,85,33],"project\u002Fneutral\u002Fidea",[107,31,33,124,11,117],"maker\u002Fneutral\u002Fidea",[56,40,40,122,57,33],[127,40,39,114,78,33],"12620843",[91,40,40,122,92,33],[130,40,33,124,11,117],"12621556",[132,40,33,133,92,33],"12638358","other\u002Fpositive\u002Fidea",[135,40,33,136,57,33],"12639934","other\u002Fneutral\u002Fidea",[],[139,143,148,153,157,162],{"id":140,"name":141,"points":142,"category":20},"12612246","Wave function collapse",1226,{"id":144,"name":145,"points":146,"category":147},"12621680","Ascii to SVG",416,"dev_tools",{"id":149,"name":150,"points":151,"category":152},"12646083","Mastodon",364,"infrastructure",{"id":154,"name":155,"points":156,"category":147},"12639612","christophercliff.com",158,{"id":158,"name":159,"points":160,"category":161},"12640851","Cypht",110,"productivity",{"id":163,"name":164,"points":165,"category":166},"12621655","Forgiva",97,"security_privacy",false,"v2",{"id":170,"name":5,"line":6,"ai":167,"cat":20,"year":13,"points":14,"comments":15,"rec":28,"s":171,"top":18,"kw":172,"maker":11,"domain":10,"tier":19,"attempts":33,"strip":173,"url":9,"mix":174,"praise":48,"objection":61,"sub":21},12619694,0.18,[24,25,26,27],"503030353000",[31,32,33,33],1794036548234]