[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-new-parser-learns-structural-rules-without-hand-coded-grammar":10,"sections":41},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":30,"tags":31,"sources":36,"feedback":40,"feedback_at":22,"cost_usd":40,"total_tokens":40},5597,"new-parser-learns-structural-rules-without-hand-coded-grammar","New Parser Learns Structural Rules Without Hand-Coded Grammar","A cellular automaton parser learns compositional rules from data alone, trailing hand-coded AM-Parser by 3.5 points on the SLOG benchmark.","A neural network just taught itself grammar rules that used to require human linguists.\n\nResearchers built a semantic parser using a neural cellular automaton (NCA) with a discrete bottleneck, meaning every compositional rule is learned from data through local iteration rather than hand-coded. Tested on the SLOG benchmark for structural generalization, the system hit 67.3% overall accuracy across 10 seeds, compared to 70.8% for AM-Parser, the existing system built on hand-written algebraic rules. The NCA actually beat AM-Parser outright on 11 of 17 structural categories, including three where AM-Parser scored between 0% and 74%. Error analysis found all 5,539 failures traced back to just two recurring patterns: novel combinations of wh-extraction with reduced verb types, and modifiers placed on the subject side of verbs.\n\nThat's the real finding here: generalization looks binary, not gradual. When researchers broke results down by CCG structural features, each sub-pattern either worked on every instance or failed on every instance. No partial credit. Scores that looked like partial generalization, such as 41.4%, turned out to be blends of totally different structural patterns scored together.\n\nTransformer-based parsers have historically struggled to generalize to new structural combinations at all, which is why hand-written systems like AM-Parser have stuck around. A learned system landing within striking distance, and beating it category by category, suggests the rules AM-Parser encodes by hand can be recovered from data given the right architecture. It's not there yet on raw accuracy, and 3.5 points still matters when the categories that fail, fail completely.","[\"semantic parsing\",\"compositional generalization\",\"neural networks\",\"nlp\"]","2026-08-18T04:00:00.000Z","2026-08-19T03:04:49.069Z","2026-08-19T03:05:00.903Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"The dek claims the neural system 'matches' AM-Parser, but the body states it trails by 3.5 points (67.3% vs 70.8%) — reword the dek so it doesn't overstate parity that the body itself contradicts.","resolved","ai",[32,33,34,35],"semantic parsing","compositional generalization","neural networks","nlp",[37],{"name":38,"url":39},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.26157",0,{"sections":42},[43,47,51,56,61,66,71,76,81,85,90,95,100,105],{"name":44,"slug":30,"count":45,"latest_published_at":46},"AI",3293,"2026-08-20T04:00:00.000Z",{"name":48,"slug":49,"count":50,"latest_published_at":46},"Security","security",435,{"name":52,"slug":53,"count":54,"latest_published_at":55},"Policy","policy",210,"2026-08-19T09:32:27.000Z",{"name":57,"slug":58,"count":59,"latest_published_at":60},"Deals","deals",179,"2026-06-29T20:02:07.000Z",{"name":62,"slug":63,"count":64,"latest_published_at":65},"Hardware","hardware",140,"2026-08-19T18:25:42.000Z",{"name":67,"slug":68,"count":69,"latest_published_at":70},"Consumer Tech","consumer-tech",95,"2026-08-18T16:05:00.000Z",{"name":72,"slug":73,"count":74,"latest_published_at":75},"Science","science",90,"2026-08-19T18:41:02.000Z",{"name":77,"slug":78,"count":79,"latest_published_at":80},"Software","software",73,"2026-08-18T07:51:50.000Z",{"name":82,"slug":83,"count":84,"latest_published_at":18},"Dev Tools","dev-tools",69,{"name":86,"slug":87,"count":88,"latest_published_at":89},"Startups","startups",47,"2026-08-19T19:13:46.000Z",{"name":91,"slug":92,"count":93,"latest_published_at":94},"Gaming","gaming",41,"2026-07-09T04:00:00.000Z",{"name":96,"slug":97,"count":98,"latest_published_at":99},"General","general",33,"2026-08-18T22:18:13.000Z",{"name":101,"slug":102,"count":103,"latest_published_at":104},"Reviews","reviews",20,"2026-06-24T12:00:01.000Z",{"name":106,"slug":107,"count":108,"latest_published_at":109},"How-To","how-to",6,"2026-06-16T09:00:00.000Z"]