mlsanityData to 14 Oct 2026

Hierarchical Response Preservation for Continual Adaptation of Zero-Shot Graph-Text Models

arXiv 2609.33607 ↗First posted 27 Sep 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Pretrained graph-text models align graph representations with textual semantics, enabling recognition of unseen classes and transfer across graph domains. However, as graph data and classes continually arrive, models should learn from new supervision while retaining their zero-shot transfer capabilities and historical task knowledge. Two challenges arise: (i) new classes can overturn historical predictions despite preserved distinctions among historical classes, and (ii) overly strict response preservation can stall learning of new classes. To address these challenges, we propose Hierarchical Response Preservation (HiRP). HiRP represents this competition through a hierarchical response that keeps each historical-class probability and sums new-class probabilities, preserving historical distinctions and aggregate competition while allowing distinctions within the new class group to adapt. It further uses the geometry induced by this response to guide constrained updates, retaining useful adaptation directions while controlling response drift. Across three class-incremental settings, HiRP achieves absolute gains of 1.84-7.95 percentage points in average accuracy over the strongest compared baseline in each setting, while mitigating zero-shot transfer degradation.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 29 of its 33 references that are in the map. First posted on arXiv 27 Sep 2026.

Builds on

29 of 33 references.

References · 29

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  3. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  4. 2016Overcoming catastrophic forgetting in neural networkscited by 4,787
  5. 2015Trust Region Policy Optimizationcited by 2,580
  6. 2016Learning without Forgettingcited by 2,496
  7. 2016iCaRL: Incremental Classifier and Representation Learningcited by 2,333
  8. 2017Memory Aware Synapses: Learning what (not) to forgetcited by 1,105
  9. 2021Learning to Prompt for Continual Learningcited by 944
  10. 2020Dark Experience for General Continual Learning: a Strong, Simple Baselinecited by 911
  11. 2022Decoupled Knowledge Distillationcited by 388
  12. 2019Orthogonal Gradient Descent for Continual Learningcited by 369
17 more
  1. 2023GraphGPT: Graph Instruction Tuning for Large Language Modelscited by 226
  2. 2024InfLoRA: Interference-Free Low-Rank Adaptation for Continual Learningcited by 198
  3. 2020Modeling the Background for Incremental Learning in Semantic Segmentationcited by 190
  4. 2023Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Needcited by 184
  5. 2024LLaGA: Large Language and Graph Assistantcited by 181
  6. 2019Functional Regularisation for Continual Learning with Gaussian Processescited by 123
  7. 2020Continual Deep Learning by Functional Regularisation of Memorable Pastcited by 105
  8. 2022TRGP: Trust Region Gradient Projection for Continual Learningcited by 82
  9. 2025SD-LoRA: Scalable Decoupled Low-Rank Adaptation for Class Incremental Learningcited by 63
  10. 2024GraphCLIP: Enhancing Transferability in Graph Foundation Models for Text-Attributed Graphscited by 58
  11. 2021Natural continual learning: success is a journey, not (just) a destinationcited by 46
  12. 2023Few-Shot Class-Incremental Learning via Training-Free Prototype Calibrationcited by 46
  13. 2024Replay-and-Forget-Free Graph Class-Incremental Learning: A Task Profiling and Prompting Approachcited by 12
  14. 2025Can LLMs Alleviate Catastrophic Forgetting in Graph Continual Learning? A Systematic Studycited by 5
  15. 2025Learning Noise-Resilient and Transferable Graph-Text Alignment via Dynamic Quality Assessmentcited by 3
  16. 2026Trust Region Continual Learning as an Implicit Meta-Learnercited by 1
  17. 2026G2LoRA: Gradient Orthogonal Low-Rank Adaptation Framework for Graph Continual Learning on Text-Attributed Graphscited by 1

Built on by

Not cited yet.