# Two Routes to the Middle: Placement Search and Brain Readouts Converge on Where Continual Learners Should Specialize

arXiv: https://arxiv.org/abs/2610.01590 · HTML: https://mlsanity.com/paper/2610.01590/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 34 of its 48 references that are in the map.
- First posted on arXiv 1 Oct 2026.

## Builds on (most cited first, 30 of 34)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Parameter-Efficient Transfer Learning for NLP](https://mlsanity.com/paper/1902.00751/) (2019)
- [The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalization](https://mlsanity.com/paper/2006.16241/) (2020)
- [How transferable are features in deep neural networks?](https://mlsanity.com/paper/1411.1792/) (2014)
- [Learning to Prompt for Continual Learning](https://mlsanity.com/paper/2112.08654/) (2021)
- [iBOT: Image BERT Pre-Training with Online Tokenizer](https://mlsanity.com/paper/2111.07832/) (2021)
- [A Comprehensive Survey of Continual Learning: Theory, Method and Application](https://mlsanity.com/paper/2302.00487/) (2023)
- [DualPrompt: Complementary Prompting for Rehearsal-free Continual Learning](https://mlsanity.com/paper/2204.04799/) (2022)
- [How to train your ViT? Data, Augmentation, and Regularization in Vision Transformers](https://mlsanity.com/paper/2106.10270/) (2021)
- [Expert Gate: Lifelong Learning with a Network of Experts](https://mlsanity.com/paper/1611.06194/) (2016)
- [Surgical Fine-Tuning Improves Adaptation to Distribution Shifts](https://mlsanity.com/paper/2210.11466/) (2022)
- [InfLoRA: Interference-Free Low-Rank Adaptation for Continual Learning](https://mlsanity.com/paper/2404.00228/) (2024)
- [Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Need](https://mlsanity.com/paper/2303.07338/) (2023)
- [FearNet: Brain-Inspired Model for Incremental Learning](https://mlsanity.com/paper/1711.10563/) (2017)
- [RanPAC: Random Projections and Pre-trained Models for Continual Learning](https://mlsanity.com/paper/2307.02251/) (2023)
- [Deep Neural Networks Reveal a Gradient in the Complexity of Neural Representations across the Brain's Ventral Visual Pathway](https://mlsanity.com/paper/1411.6422/) (2014)
- [First Session Adaptation: A Strong Replay-Free Baseline for Class-Incremental Learning](https://mlsanity.com/paper/2303.13199/) (2023)
- [Inducing brain-relevant bias in natural language processing models](https://mlsanity.com/paper/1911.03268/) (2019)
- [Learning From Brains How to Regularize Machines](https://mlsanity.com/paper/1911.05072/) (2019)
- [LoRA-drop: Efficient LoRA Parameter Pruning based on Output Evaluation](https://mlsanity.com/paper/2402.07721/) (2024)
- [HiDe-LLaVA: Hierarchical Decoupling for Continual Instruction Tuning of Multimodal Large Language Model](https://mlsanity.com/paper/2503.12941/) (2025)
- [CL-LoRA: Continual Low-Rank Adaptation for Rehearsal-Free Class-Incremental Learning](https://mlsanity.com/paper/2505.24816/) (2025)
- [Dynamic Mixture of Curriculum LoRA Experts for Continual Multimodal Instruction Tuning](https://mlsanity.com/paper/2506.11672/) (2025)
- [Improving Semantic Understanding in Speech Language Models via Brain-tuning](https://mlsanity.com/paper/2410.09230/) (2024)
- [Brain Decodes Deep Nets](https://mlsanity.com/paper/2312.01280/) (2023)
- [TreeLoRA: Efficient Continual Learning via Layer-Wise LoRAs Guided by a Hierarchical Gradient-Similarity Tree](https://mlsanity.com/paper/2506.10355/) (2025)
- [RainbowPrompt: Diversity-Enhanced Prompt-Evolving for Continual Learning](https://mlsanity.com/paper/2507.22553/) (2025)
- [BrainWavLM: Fine-tuning Speech Representations with Brain Responses to Language](https://mlsanity.com/paper/2502.08866/) (2025)
- [Brain-tuning Improves Generalizability and Efficiency of Brain Alignment in Speech Models](https://mlsanity.com/paper/2510.21520/) (2025)

## Abstract (arXiv)

Continual learners that keep a task-specific adapter in every block of a pre-trained vision transformer accumulate storage linearly with the number of tasks; keeping task-specific adapters in only a few blocks curbs this growth but raises the question of where to place them. We investigate this question from two perspectives. Algorithmically, training all contiguous four-block placements yields an inverted U: final accuracy peaks at intermediate depth and varies by up to 3.5 percentage points (pp), while inexpensive criteria based on weight spectra or activation statistics favor the deepest blocks. From neuroscience, the hierarchical organization and intermediate-stage plasticity of the visual cortex motivate us to ask whether a measurement taken outside the learner can guide layer specialization without placement search. LS-B observes the first tasks through a frozen fMRI encoding model of twelve human visual areas and commits task-specific capacity once to the blocks whose readouts vary most across tasks relative to their stable structure. Across three ViT-B/16 backbones, LS-B yields stable, backbone-specific allocations. On the two backbones with placement search, AugReg and iBOT, the selected blocks overlap the intermediate-depth region identified by search. Under matched storage and observation budgets, the selected blocks outperform the shallowest and deepest four-block configurations. On Split ImageNet-R, LS-B uses 60% of full-BiLoRA adapter storage while remaining within 1.5 pp of its final accuracy. The allocation requires no labels or backpropagation, adds under 0.6% runtime, and exhibits backbone-specific cortical signatures.