mlsanityData to 7 Oct 2026

SSH: Sparse Spectrum Adaptation via Discrete Hartley Transformation

arXiv 2502.05539 ↗First posted 8 Feb 2025cs.CV cs.LGcited by 7

Not on the map. Why?

Abstract (arXiv)

Low-rank adaptation (LoRA) has been demonstrated effective in reducing the trainable parameter number when fine-tuning a large foundation model (LLM). However, it still encounters computational and memory challenges when scaling to larger models or addressing more complex task adaptation. In this work, we introduce Sparse Spectrum Adaptation via Discrete Hartley Transformation (SSH), a novel approach that significantly reduces the number of trainable parameters while enhancing model performance. It selects the most informative spectral components across all layers, under the guidance of the initial weights after a discrete Hartley transformation (DHT). The lightweight inverse DHT then projects the spectrum back into the spatial domain for updates. Extensive experiments across both single-modality tasks such as language understanding and generation and multi-modality tasks such as video-text understanding demonstrate that SSH outperforms existing parameter-efficient fine-tuning (PEFT) methods while achieving substantial reductions in computational cost and memory requirements.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Builds on 37 of its 59 references that are in the map. First posted on arXiv 8 Feb 2025.

Builds on

37 of 59 references.

References · 37

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  3. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  4. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  5. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  6. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  7. 2019BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehensioncited by 5,874
  8. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  9. 2023Improved Baselines with Visual Instruction Tuningcited by 4,362
  10. 2019Parameter-Efficient Transfer Learning for NLPcited by 4,195
  11. 2023QLoRA: Efficient Finetuning of Quantized LLMscited by 3,355
  12. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,488
25 more
  1. 2018Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarizationcited by 1,308
  2. 2015Teaching Machines to Read and Comprehendcited by 1,039
  3. 2021Towards a Unified View of Parameter-Efficient Transfer Learningcited by 863
  4. 2020AdapterFusion: Non-Destructive Task Composition for Transfer Learningcited by 780
  5. 2024DoRA: Weight-Decomposed Low-Rank Adaptationcited by 740
  6. 2024Sora: A Review on Background, Technology, Limitations, and Opportunities of Large Vision Modelscited by 467
  7. 20218-bit Optimizers via Block-wise Quantizationcited by 395
  8. 2023VeRA: Vector-based Random Matrix Adaptationcited by 294
  9. 2017The E2E Dataset: New Challenges For End-to-End Generationcited by 252
  10. 2020AdapterDrop: On the Efficiency of Adapters in Transformerscited by 223
  11. 2023ReLoRA: High-Rank Training Through Low-Rank Updatescited by 200
  12. 2020Learning in the Frequency Domaincited by 196
  13. 2022Rethinking Graph Neural Networks for Anomaly Detectioncited by 185
  14. 2020Exploring Versatile Generative Language Model Via Parameter-Efficient Transfer Learningcited by 139
  15. 2023Tied-Lora: Enhancing parameter efficiency of LoRA with weight tyingcited by 65
  16. 2023Parameter-Efficient Fine-Tuning without Introducing New Latencycited by 59
  17. 2023Make Pre-trained Model Reversible: From Parameter to Memory Efficient Fine-Tuningcited by 26
  18. 2024Learning Spectral-Decomposed Tokens for Domain Generalized Semantic Segmentationcited by 21
  19. 2024LaMDA: Large Model Fine-Tuning via Spectrally Decomposed Low-Dimensional Adaptationcited by 16
  20. 2024AFLoRA: Adaptive Freezing of Low Rank Adaptation in Parameter Efficient Fine-Tuning of Large Modelscited by 14
  21. 2024A Novel Evaluation Framework for Image2Text Generationcited by 11
  22. 2024Gradient Weight-normalized Low-rank Projection for Efficient LLM Trainingcited by 11
  23. 2024AltGen: AI-Driven Alt Text Generation for Enhancing EPUB Accessibilitycited by 11
  24. 2024Comparative Analysis of Listwise Reranking with Large Language Models in Limited-Resource Language Contextscited by 7
  25. 2022Low-complexity Three-dimensional Discrete Hartley Transform Approximations for Medical Image Compressioncited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 3; 2026: 4.

7 papers.

  1. 2025DeepEyeNet: Generating Medical Report for Retinal Imagescited by 1
  2. 2025QWHA: Quantization-Aware Walsh-Hadamard Adaptation for Parameter-Efficient Fine-Tuning on Large Language Modelscited by 1
  3. 2026Spectral-Progressive Thought Flow for Lightweight Multimodal Reasoningcited by 1
  4. 2025MaCP: Minimal yet Mighty Adaptation via Hierarchical Cosine Projectioncited by 0
  5. 2026A Novel Automatic Framework for Speaker Drift Detection in Synthesized Speechcited by 0
  6. 2026Active Imitation Learning for Thermal- and Kernel-Aware LFM Inference on 3D S-NUCA Many-Corescited by 0
  7. 2026TCL: Enabling Fast and Efficient Cross-Hardware Tensor Program Optimization via Continual Learningcited by 0