{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:NBGSRIW6WQJGDWQRLUZTKAQ3LC","short_pith_number":"pith:NBGSRIW6","canonical_record":{"source":{"id":"2604.12110","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-13T22:33:28Z","cross_cats_sorted":[],"title_canon_sha256":"fbf2384f5061344f90dbf2221e0a38bd0fc384c2805749062608c9412800081a","abstract_canon_sha256":"ffd13573684688b04d111fa68173180411b02a19e90ca2c9a5422c8966a871e9"},"schema_version":"1.0"},"canonical_sha256":"684d28a2deb41261da115d3335021b588fe39c2786321f4d0d82777446f541ca","source":{"kind":"arxiv","id":"2604.12110","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.12110","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"arxiv_version","alias_value":"2604.12110v2","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.12110","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_12","alias_value":"NBGSRIW6WQJG","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_16","alias_value":"NBGSRIW6WQJGDWQR","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_8","alias_value":"NBGSRIW6","created_at":"2026-06-05T01:14:38Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:NBGSRIW6WQJGDWQRLUZTKAQ3LC","target":"record","payload":{"canonical_record":{"source":{"id":"2604.12110","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-13T22:33:28Z","cross_cats_sorted":[],"title_canon_sha256":"fbf2384f5061344f90dbf2221e0a38bd0fc384c2805749062608c9412800081a","abstract_canon_sha256":"ffd13573684688b04d111fa68173180411b02a19e90ca2c9a5422c8966a871e9"},"schema_version":"1.0"},"canonical_sha256":"684d28a2deb41261da115d3335021b588fe39c2786321f4d0d82777446f541ca","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:14:38.481913Z","signature_b64":"gs6Ga3WxUebHTlDjXIJ45GmeciQ81JnpzoljclrBnfAL8fxAGjBTF08amUIhnMvvRKLHllc33xVt26TsnpTQAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"684d28a2deb41261da115d3335021b588fe39c2786321f4d0d82777446f541ca","last_reissued_at":"2026-06-05T01:14:38.481258Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:14:38.481258Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.12110","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-05T01:14:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"lwFiF+ROmU1OB0dnEQ9GM18HvfxWQwg9AJ/gscttKcemQsViI5xX0/9rFrRsyE3cTS9Rr6NxIRzCvjTzmdHhDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-28T10:31:54.456222Z"},"content_sha256":"b2237ec35194e4f32a1a1ccd4a5ea36089421feee64cfe89e02b5b236da72ea1","schema_version":"1.0","event_id":"sha256:b2237ec35194e4f32a1a1ccd4a5ea36089421feee64cfe89e02b5b236da72ea1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:NBGSRIW6WQJGDWQRLUZTKAQ3LC","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"SOLARIS: Speculative Offloading of Latent-bAsed Representation for Inference Scaling","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"Predicting future user-item pairs allows precomputing their embeddings to use complex foundation models in real-time serving.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Abdallah Aboelela, Chonglin Sun, Dong Liang, Ellie Wen, Feifan Gu, Fenggang Wu, Hang Qu, Huayu Li, Jill Pan, Jingxian Huang, Jingyi Shen, Jing Zhu, Kaidi Pei, Laming Chen, Liang Luo, Longhao Jin, Qianru Li, Qin Huang, Tongyi Tang, Varna Puvvada, Wei Ling, Wenlin Chen, Xiaohan Wei, Xu Cao, Yaning Huang, Yantao Yao, Yuan Jin, Yunchen Pu, Yuxin Chen, Zeliang Chen, Zhengkai Zhang, Zhengyu Zhang, Zijian Shen, Zikun Liu","submitted_at":"2026-04-13T22:33:28Z","abstract_excerpt":"Recent advances in recommendation scaling laws have led to foundation models of unprecedented complexity. While these models offer superior performance, their computational demands make real-time serving impractical, often forcing practitioners to rely on knowledge distillation-compromising serving quality for efficiency. To address this challenge, we present SOLARIS (Speculative Offloading of Latent-bAsed Representation for Inference Scaling), a novel framework inspired by speculative decoding. SOLARIS proactively precomputes user-item interaction embeddings by predicting which user-item pair"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Deployed across Meta's advertising system serving billions of daily requests, SOLARIS achieves 0.67% revenue-driving top-line metrics gain, demonstrating its effectiveness at scale.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That accurate prediction of which user-item pairs will appear in future requests is feasible at low enough error rate to deliver net positive value after accounting for precomputation cost and any stale or unused embeddings.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SOLARIS speculatively precomputes user-item latent representations to decouple large-model inference from real-time serving, delivering 0.67% revenue gain when deployed in Meta's ad system.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Predicting future user-item pairs allows precomputing their embeddings to use complex foundation models in real-time serving.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"fca57890cea0b308f781aa86aa8aa3a25cc1a3dd58f6f508cf6a82234ab31778"},"source":{"id":"2604.12110","kind":"arxiv","version":2},"verdict":{"id":"bf0a4ad5-1a34-465c-b2a0-4c9f4e07bba5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T15:01:26.487506Z","strongest_claim":"Deployed across Meta's advertising system serving billions of daily requests, SOLARIS achieves 0.67% revenue-driving top-line metrics gain, demonstrating its effectiveness at scale.","one_line_summary":"SOLARIS speculatively precomputes user-item latent representations to decouple large-model inference from real-time serving, delivering 0.67% revenue gain when deployed in Meta's ad system.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That accurate prediction of which user-item pairs will appear in future requests is feasible at low enough error rate to deliver net positive value after accounting for precomputation cost and any stale or unused embeddings.","pith_extraction_headline":"Predicting future user-item pairs allows precomputing their embeddings to use complex foundation models in real-time serving."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.12110/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"bf0a4ad5-1a34-465c-b2a0-4c9f4e07bba5"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-05T01:14:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oDr/L9y09w6qPNjvF2szw8yd9rh7/E1k2RzS+Skkf+mOxQUrFuGcZpMYCJcqHlztY2yIDpSPK5lTctkLDQDHCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-28T10:31:54.456677Z"},"content_sha256":"aebcdbf12fc26f4161ffb731373b61bb9c952a704f985c19319a51a423001079","schema_version":"1.0","event_id":"sha256:aebcdbf12fc26f4161ffb731373b61bb9c952a704f985c19319a51a423001079"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/NBGSRIW6WQJGDWQRLUZTKAQ3LC/bundle.json","state_url":"https://pith.science/pith/NBGSRIW6WQJGDWQRLUZTKAQ3LC/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/NBGSRIW6WQJGDWQRLUZTKAQ3LC/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-28T10:31:54Z","links":{"resolver":"https://pith.science/pith/NBGSRIW6WQJGDWQRLUZTKAQ3LC","bundle":"https://pith.science/pith/NBGSRIW6WQJGDWQRLUZTKAQ3LC/bundle.json","state":"https://pith.science/pith/NBGSRIW6WQJGDWQRLUZTKAQ3LC/state.json","well_known_bundle":"https://pith.science/.well-known/pith/NBGSRIW6WQJGDWQRLUZTKAQ3LC/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:NBGSRIW6WQJGDWQRLUZTKAQ3LC","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"ffd13573684688b04d111fa68173180411b02a19e90ca2c9a5422c8966a871e9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-13T22:33:28Z","title_canon_sha256":"fbf2384f5061344f90dbf2221e0a38bd0fc384c2805749062608c9412800081a"},"schema_version":"1.0","source":{"id":"2604.12110","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.12110","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"arxiv_version","alias_value":"2604.12110v2","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.12110","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_12","alias_value":"NBGSRIW6WQJG","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_16","alias_value":"NBGSRIW6WQJGDWQR","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_8","alias_value":"NBGSRIW6","created_at":"2026-06-05T01:14:38Z"}],"graph_snapshots":[{"event_id":"sha256:aebcdbf12fc26f4161ffb731373b61bb9c952a704f985c19319a51a423001079","target":"graph","created_at":"2026-06-05T01:14:38Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Deployed across Meta's advertising system serving billions of daily requests, SOLARIS achieves 0.67% revenue-driving top-line metrics gain, demonstrating its effectiveness at scale."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That accurate prediction of which user-item pairs will appear in future requests is feasible at low enough error rate to deliver net positive value after accounting for precomputation cost and any stale or unused embeddings."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"SOLARIS speculatively precomputes user-item latent representations to decouple large-model inference from real-time serving, delivering 0.67% revenue gain when deployed in Meta's ad system."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Predicting future user-item pairs allows precomputing their embeddings to use complex foundation models in real-time serving."}],"snapshot_sha256":"fca57890cea0b308f781aa86aa8aa3a25cc1a3dd58f6f508cf6a82234ab31778"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.12110/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advances in recommendation scaling laws have led to foundation models of unprecedented complexity. While these models offer superior performance, their computational demands make real-time serving impractical, often forcing practitioners to rely on knowledge distillation-compromising serving quality for efficiency. To address this challenge, we present SOLARIS (Speculative Offloading of Latent-bAsed Representation for Inference Scaling), a novel framework inspired by speculative decoding. SOLARIS proactively precomputes user-item interaction embeddings by predicting which user-item pair","authors_text":"Abdallah Aboelela, Chonglin Sun, Dong Liang, Ellie Wen, Feifan Gu, Fenggang Wu, Hang Qu, Huayu Li, Jill Pan, Jingxian Huang, Jingyi Shen, Jing Zhu, Kaidi Pei, Laming Chen, Liang Luo, Longhao Jin, Qianru Li, Qin Huang, Tongyi Tang, Varna Puvvada, Wei Ling, Wenlin Chen, Xiaohan Wei, Xu Cao, Yaning Huang, Yantao Yao, Yuan Jin, Yunchen Pu, Yuxin Chen, Zeliang Chen, Zhengkai Zhang, Zhengyu Zhang, Zijian Shen, Zikun Liu","cross_cats":[],"headline":"Predicting future user-item pairs allows precomputing their embeddings to use complex foundation models in real-time serving.","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-13T22:33:28Z","title":"SOLARIS: Speculative Offloading of Latent-bAsed Representation for Inference Scaling"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.12110","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-10T15:01:26.487506Z","id":"bf0a4ad5-1a34-465c-b2a0-4c9f4e07bba5","model_set":{"reader":"grok-4.3"},"one_line_summary":"SOLARIS speculatively precomputes user-item latent representations to decouple large-model inference from real-time serving, delivering 0.67% revenue gain when deployed in Meta's ad system.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Predicting future user-item pairs allows precomputing their embeddings to use complex foundation models in real-time serving.","strongest_claim":"Deployed across Meta's advertising system serving billions of daily requests, SOLARIS achieves 0.67% revenue-driving top-line metrics gain, demonstrating its effectiveness at scale.","weakest_assumption":"That accurate prediction of which user-item pairs will appear in future requests is feasible at low enough error rate to deliver net positive value after accounting for precomputation cost and any stale or unused embeddings."}},"verdict_id":"bf0a4ad5-1a34-465c-b2a0-4c9f4e07bba5"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:b2237ec35194e4f32a1a1ccd4a5ea36089421feee64cfe89e02b5b236da72ea1","target":"record","created_at":"2026-06-05T01:14:38Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"ffd13573684688b04d111fa68173180411b02a19e90ca2c9a5422c8966a871e9","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2026-04-13T22:33:28Z","title_canon_sha256":"fbf2384f5061344f90dbf2221e0a38bd0fc384c2805749062608c9412800081a"},"schema_version":"1.0","source":{"id":"2604.12110","kind":"arxiv","version":2}},"canonical_sha256":"684d28a2deb41261da115d3335021b588fe39c2786321f4d0d82777446f541ca","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"684d28a2deb41261da115d3335021b588fe39c2786321f4d0d82777446f541ca","first_computed_at":"2026-06-05T01:14:38.481258Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-05T01:14:38.481258Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"gs6Ga3WxUebHTlDjXIJ45GmeciQ81JnpzoljclrBnfAL8fxAGjBTF08amUIhnMvvRKLHllc33xVt26TsnpTQAQ==","signature_status":"signed_v1","signed_at":"2026-06-05T01:14:38.481913Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.12110","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b2237ec35194e4f32a1a1ccd4a5ea36089421feee64cfe89e02b5b236da72ea1","sha256:aebcdbf12fc26f4161ffb731373b61bb9c952a704f985c19319a51a423001079"],"state_sha256":"3f4730304f2622ade6768c99f79b4b077f510d246b98d37f58433c655e1737b7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2TdqhEay/669Spj/HltcSBjXHpui0y+jBYhYeQzR+4onZVSK2vUEAYANPfHre7QfHWeuj/1P0KQou3+ATupjBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-28T10:31:54.458843Z","bundle_sha256":"07781920dac5f969ce428ce1fe806f05e3291c878ebd3d30ae5facbfa10f3296"}}