ARXIV:2601.19142 · AI-POWERED RECOMMENDER SYSTEMS · SUBMITTED 19 MAR · 18:48 UTC · FRESHNESS STALE

VerifiedSource: PDF linkedPartialPaperPack: 3 of 4 citation fields filledMissingMissing fields: authorsPartialProof: unverified proof status

Length-Adaptive Interest Network for Balancing Long and Short Sequence Modeling in CTR Prediction

arXiv

Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network.

Blocked on Code›Score7.0Evidence unverified

Opportunity summary

Pain Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network.

Evidence 0 refs | 0 sources | 33% coverage

Blocker Evidence unverified

Open Build Read PDF Signal Canvas Track

PROBLEM

Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network. While longer sequences provide more context, we observe that increasing the maximum input sequence length in existing CTR models…

METHOD

Full abstract

User behavior sequences in modern recommendation systems exhibit significant length heterogeneity, ranging from sparse short-term interactions to rich long-term histories. While longer sequences provide more context, we observe that increasing the maximum input sequence length in existing CTR models paradoxically degrades performance for short-sequence users due to attention polarization and length imbalance in training data. To address this, we propose LAIN(Length-Adaptive Interest Network), a plug-and-play framework that explicitly incorporates sequence length as a conditioning signal to balance long- and short-sequence modeling. LAIN consists of three lightweight components: a Spectral Length Encoder that maps length into continuous representations, Length-Conditioned Prompting that injects global contextual cues into both long- and short-term behavior branches, and Length-Modulated Attention that adaptively adjusts attention sharpness based on sequence length. Extensive experiments on three real-world benchmarks across five strong CTR backbones show that LAIN consistently improves overall performance, achieving up to 1.15% AUC gain and 2.25% log loss reduction. Notably, our method significantly improves accuracy for short-sequence users without sacrificing longsequence effectiveness. Our work offers a general, efficient, and deployable solution to mitigate length-induced bias in sequential recommendation.

RESULT

ScienceToStartup currently rates this 7.0/10 on the public viability pass. Extensive experiments on three real-world benchmarks across five strong CTR backbones show that LAIN consistently improves overall performance, achieving up to 1.15% AUC gain…

WHY NOW

AI-Powered Recommender Systems moved forward this cycle; last verified April 2026. Public score 7.0/10.

Continue into Read for claims, analysis, references, and neighboring papers.

Opportunity summary

Score7.0

PainEnhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network.

Evidence0 refs | 0 sources | 33% coverage

Blockermissing authors

Analysis summary

Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network.

VerifiedSource: PDF linkedPartialPaperPack: 3 of 4 citation fields filledMissingMissing fields: authorsPartialProof: unverified proof status

Competitive landscape

Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network.

Segment

AI-Powered Recommender Systems

Adoption evidence

No public code link in the paper record yet

Commercial read

7.0/10 public viability

Direct

not classified

Adjacent

not classified

Substitute

not classified

Unknown

not classified

References(16)

GMSA: Enhancing Context Compression via Group Merging and Layer Semantic Alignment

2025Jiwei Tang, Zhicheng Zhang et al.

Long-Sequence Recommendation Models Need Decoupled Embeddings

2024Ningya Feng, Junwei Pan et al.

Perception Compressor: A Training-Free Prompt Compression Framework in Long Context Scenarios

2024Jiwei Tang, Jin Xu et al.

Prompt Tuning as User Inherent Profile Inference Machine

2024Yusheng Lu, Zhaocheng Du et al.

Efficient Noise-Decoupling for Multi-Behavior Sequential Recommendation

2024Yongqiang Han, Hao Wang et al.

Improving Long-Tail Item Recommendation with Graph Augmentation

2023Sichun Luo, Chen Ma et al.

FinalMLP: An Enhanced Two-Stream MLP Model for CTR Prediction

2023Kelong Mao, Jieming Zhu et al.

BARS: Towards Open Benchmarking for Recommender Systems

2022Jieming Zhu, Kelong Mao et al.

Empowering General-purpose User Representation with Full-life Cycle Behavior Modeling

2021Bei Yang, Jie Gu et al.

Learning Transferrable Parameters for Long-tailed Sequential User Behavior Modeling

2020Jianwen Yin, Chenghao Liu et al.

Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains

2020Matthew Tancik, Pratul P. Srinivasan et al.

Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Prediction

2020Qi Pi, Xiaoqiang Zhu et al.

MeLU: Meta-Learned User Preference Estimator for Cold-Start Recommendation

2019Hoyeop Lee, Jinbae Im et al.

Temporal Hierarchical Attention at Category- and Item-Level for Micro-Video Click-Through Prediction

2018Xusong Chen, Dong Liu et al.

Deep Interest Evolution Network for Click-Through Rate Prediction

2018Guorui Zhou, Na Mou et al.

Deep Interest Network for Click-Through Rate Prediction

2017Guorui Zhou, Cheng-Ning Song et al.

{ "contract_version": "paper-r2", "paper_id": "0694d679-a4dd-4ac7-9e1b-3617ab80c224", "arxiv_id": "2601.19142", "canonical_route": "/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction", "active_tab": "synced from current hash by the drawer client", "selected_artifact": "length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction", "endpoints": { "paper_pack": "/api/v1/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction/paper-pack", "build_passport": "/api/v1/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction/build-passport", "mcp_resource": "sciencetostartup://surfaces/paper-workspace" } }

{ "surface": "paper", "mode": "paper", "query": "Length-Adaptive Interest Network for Balancing Long and Short Sequence Modeling in CTR Prediction", "normalized_query": "2601.19142", "route": "/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction", "paper_ref": "length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction", "topic_slug": null, "benchmark_ref": null, "dataset_ref": null }

{ "@context": "https://schema.org", "@graph": [ { "@type": "WebPage", "@id": "https://sciencetostartup.com/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction#webpage", "url": "https://sciencetostartup.com/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction", "name": "Length-Adaptive Interest Network for Balancing Long and Short Sequence Modeling in CTR Prediction", "description": "Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network.", "isPartOf": { "@id": "https://sciencetostartup.com/#website" } }, { "@type": "ScholarlyArticle", "@id": "https://sciencetostartup.com/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction#scholarlyArticle", "headline": "Length-Adaptive Interest Network for Balancing Long and Short Sequence Modeling in CTR Prediction", "description": "Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network.", "url": "https://sciencetostartup.com/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction", "sameAs": "https://arxiv.org/abs/2601.19142", "identifier": { "@type": "PropertyValue", "propertyID": "arXiv", "value": "2601.19142" }, "isAccessibleForFree": true, "isPartOf": { "@id": "https://sciencetostartup.com/#website" }, "datePublished": "2026-01-27T03:14:20.000Z", "author": [ { "@type": "Person", "name": "Zhicheng Zhang" }, { "@type": "Person", "name": "Zhaocheng Du" }, { "@type": "Person", "name": "Jieming Zhu" }, { "@type": "Person", "name": "Jiwei Tang" }, { "@type": "Person", "name": "Fengyuan Lu" }, { "@type": "Person", "name": "Wang Jiaheng" }, { "@type": "Person", "name": "Song-Li Wu" }, { "@type": "Person", "name": "Qianhui Zhu" }, { "@type": "Person", "name": "Jingyu Li" }, { "@type": "Person", "name": "Hai-Tao Zheng" }, { "@type": "Person", "name": "Zhenhua Dong" } ], "citation": [ { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "fb8891088089924708d1a8d094baceef278cd9bb" }, "url": "https://www.semanticscholar.org/paper/fb8891088089924708d1a8d094baceef278cd9bb" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "c464ad1d4ec99c3ed8da02217df61dace3075f67" }, "url": "https://www.semanticscholar.org/paper/c464ad1d4ec99c3ed8da02217df61dace3075f67" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "544718be71f1598d95744f30498368d16f39a3e3" }, "url": "https://www.semanticscholar.org/paper/544718be71f1598d95744f30498368d16f39a3e3" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "2bc0e1b782ec7ff56bcd0d117e553f411c7940b0" }, "url": "https://www.semanticscholar.org/paper/2bc0e1b782ec7ff56bcd0d117e553f411c7940b0" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "6561212f9572d5201f2d657063cccc09ca519412" }, "url": "https://www.semanticscholar.org/paper/6561212f9572d5201f2d657063cccc09ca519412" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "7ff2c76b84b82e2d70aed50d4dc22bca6f2b5017" }, "url": "https://www.semanticscholar.org/paper/7ff2c76b84b82e2d70aed50d4dc22bca6f2b5017" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "4af58fc20efaa3856df8609921b6a022f8f9d3ac" }, "url": "https://www.semanticscholar.org/paper/4af58fc20efaa3856df8609921b6a022f8f9d3ac" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "eb217e5e245305e3cb628f5a2e9b6d0a25d1032d" }, "url": "https://www.semanticscholar.org/paper/eb217e5e245305e3cb628f5a2e9b6d0a25d1032d" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "373a080c5a3f4daa9327b819fb5919f0de5dc811" }, "url": "https://www.semanticscholar.org/paper/373a080c5a3f4daa9327b819fb5919f0de5dc811" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "f6095d04c537776c36073a18a2a5f580f402970d" }, "url": "https://www.semanticscholar.org/paper/f6095d04c537776c36073a18a2a5f580f402970d" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "a0dc3135c40e150f0271002a96b7c9680b6cac40" }, "url": "https://www.semanticscholar.org/paper/a0dc3135c40e150f0271002a96b7c9680b6cac40" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "75a618a86b67b8b32224b77eae8ed37c536826da" }, "url": "https://www.semanticscholar.org/paper/75a618a86b67b8b32224b77eae8ed37c536826da" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "27d2e30bb3ddbed3023340a5ef4d25d4a10bcc70" }, "url": "https://www.semanticscholar.org/paper/27d2e30bb3ddbed3023340a5ef4d25d4a10bcc70" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "7b82803010602356e63c0b116af1578b1a72e759" }, "url": "https://www.semanticscholar.org/paper/7b82803010602356e63c0b116af1578b1a72e759" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "70d5ceb59118334e1a6eed33a149234413147b92" }, "url": "https://www.semanticscholar.org/paper/70d5ceb59118334e1a6eed33a149234413147b92" }, { "@type": "ScholarlyArticle", "identifier": { "@type": "PropertyValue", "propertyID": "SemanticScholar", "value": "b8c6ccd5c1eb4f9837fc4877d27e55b7349781be" }, "url": "https://www.semanticscholar.org/paper/b8c6ccd5c1eb4f9837fc4877d27e55b7349781be" } ], "additionalProperty": [ { "@type": "PropertyValue", "propertyID": "viabilityScore", "value": 7 }, { "@type": "PropertyValue", "propertyID": "researchDomain", "value": "AI-Powered Recommender Systems" } ] }, { "@type": "BreadcrumbList", "itemListElement": [ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://sciencetostartup.com" }, { "@type": "ListItem", "position": 2, "name": "AI-Powered Recommender Systems", "item": "https://sciencetostartup.com/topics" }, { "@type": "ListItem", "position": 3, "name": "Length-Adaptive Interest Network for Balancing Long and Shor", "item": "https://sciencetostartup.com/paper/length-adaptive-interest-network-for-balancing-long-and-short-sequence-modeling-in-ctr-prediction" } ] }, { "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "What is the startup potential of \"Length-Adaptive Interest Network for Balancing Long and Shor\"?", "acceptedAnswer": { "@type": "Answer", "text": "Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network." } }, { "@type": "Question", "name": "What products could be built from this research?", "acceptedAnswer": { "@type": "Answer", "text": "Productize LAIN as a modular plugin for existing recommendation systems that can be easily integrated into platforms to improve user engagement metrics." } }, { "@type": "Question", "name": "What are the practical use cases?", "acceptedAnswer": { "@type": "Answer", "text": "Integrate LAIN into e-commerce or content-streaming platforms to enhance their recommendation engines, improving user engagement and personalized content delivery." } }, { "@type": "Question", "name": "What industries could this research disrupt?", "acceptedAnswer": { "@type": "Answer", "text": "LAIN offers a significant improvement over existing CTR models by addressing sequence length imbalance, potentially setting a new standard in recommendation accuracy." } } ] } ] }

Competitive landscape

Enhance CTR prediction by balancing long and short sequence modeling using a Length-Adaptive Interest Network.

Segment

AI-Powered Recommender Systems

Adoption evidence

No public code link in the paper record yet

Commercial read

7.0/10 public viability

Direct

not classified

Adjacent

not classified

Substitute

not classified

Unknown

not classified

References(16)

GMSA: Enhancing Context Compression via Group Merging and Layer Semantic Alignment

2025Jiwei Tang, Zhicheng Zhang et al.

Long-Sequence Recommendation Models Need Decoupled Embeddings

2024Ningya Feng, Junwei Pan et al.

Perception Compressor: A Training-Free Prompt Compression Framework in Long Context Scenarios

2024Jiwei Tang, Jin Xu et al.

Prompt Tuning as User Inherent Profile Inference Machine

2024Yusheng Lu, Zhaocheng Du et al.

Efficient Noise-Decoupling for Multi-Behavior Sequential Recommendation

2024Yongqiang Han, Hao Wang et al.

Improving Long-Tail Item Recommendation with Graph Augmentation

2023Sichun Luo, Chen Ma et al.

FinalMLP: An Enhanced Two-Stream MLP Model for CTR Prediction

2023Kelong Mao, Jieming Zhu et al.

BARS: Towards Open Benchmarking for Recommender Systems

2022Jieming Zhu, Kelong Mao et al.

Empowering General-purpose User Representation with Full-life Cycle Behavior Modeling

2021Bei Yang, Jie Gu et al.

Learning Transferrable Parameters for Long-tailed Sequential User Behavior Modeling

2020Jianwen Yin, Chenghao Liu et al.

Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains

2020Matthew Tancik, Pratul P. Srinivasan et al.

Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Prediction

2020Qi Pi, Xiaoqiang Zhu et al.

MeLU: Meta-Learned User Preference Estimator for Cold-Start Recommendation

2019Hoyeop Lee, Jinbae Im et al.

Temporal Hierarchical Attention at Category- and Item-Level for Micro-Video Click-Through Prediction

2018Xusong Chen, Dong Liu et al.

Deep Interest Evolution Network for Click-Through Rate Prediction

2018Guorui Zhou, Na Mou et al.

Deep Interest Network for Click-Through Rate Prediction

2017Guorui Zhou, Cheng-Ning Song et al.

Length-Adaptive Interest Network for Balancing Long and Short Sequence Modeling in CTR Prediction

Length-Adaptive Interest Network for Balancing Long and Short Sequence Modeling in CTR Prediction

Claim map

Constellation map

Competitive landscape

Buzz

PDF

References(16)

Related Papers

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline

Claim map

Constellation map

Competitive landscape

Buzz

PDF

References(16)

Related Papers

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline