paper-with-me

홈 › Papers

SensLI: Sensitivity-Based Layer Insertion for Neural Networks

2023-11-27 · Leonie Kreis, Evelyn Herberg, Frederik Köhne, Anton Schiela, Roland Herzog

The training of neural networks requires tedious and often manual tuning of the network architecture. We propose a systematic approach to inserting new layers during the training process. Our method eliminates the need to choose a fixed network size before training, is numerically inexpensive to execute and applicable to various architectures including fully connected feedforward networks, ResNets and CNNs. Our technique borrows ideas from constrained optimization and is based on first-order sensitivity information of the loss function with respect to the virtual parameters that additional layers, if inserted, would offer. In numerical experiments, our proposed sensitivity-based layer insertion technique (SensLI) exhibits improved performance on training loss and test error, compared to training on a fixed architecture, and reduced computational effort in comparison to training the extended architecture from the beginning. Our code is available on https://github.com/mathemml/SensLI.

📄 PDF Abstract BibTeX arXiv:2311.15995

Code (2)

leoniekreis/layer_insertion_sensitivity_based 공식 구현 pytorch
mathemml/sensli 공식 구현 pytorch

Tasks

Sensitivity

Similar Papers 제목 키워드 기반

Explicit Layer Modeling for Video Object Insertion and Layer Decomposition

2026-07-28 · Kyujin Han, Seungjoo Shin, Sunghyun Cho arxiv

Most video editing systems still lack explicit layered video representations, limiting their ability to perform realistic compositing, object reuse, and consistent manipulation. This limitation is especially pronounced i…

Beyond Superficial Forgetting: Thorough Unlearning through Knowledge Density Estimation and Block Re-insertion

2025-11-11 · Feng Guo, Yuntao Wen, Shen Gao, Junshuo Zhang 외 arxiv

Machine unlearning, which selectively removes harmful knowledge from a pre-trained model without retraining from scratch, is crucial for addressing privacy, regulatory compliance, and ethical concerns in Large Language M…

Density Estimation

Transformer Field Theory: A Response-Theoretic Approach to Mechanistic Interpretability

2026-05-24 · David N. Olivieri, Antonio F. Pérez Rodríguez arxiv

Mechanistic interpretability often studies Transformer behavior by intervening on internal activations through activation patching, causal tracing, path patching, and steering directions. This paper develops Transformer …

Where and What Matters: Sensitivity-Aware Task Vectors for Many-Shot Multimodal In-Context Learning

2025-11-11 · Ziyu Ma, Chenhui Gou, Yiming Hu, Yong Wang 외 arxiv

Large Multimodal Models (LMMs) have shown promising in-context learning (ICL) capabilities, but scaling to many-shot settings remains difficult due to limited context length and high inference cost. To address these chal…

Reinforcement Learning

A hybrid P3HT-Graphene interface for efficient photostimulation of neurons

2020-03-25 · Mattia L. DiFrancesco, Elisabetta Colombo, Ermanno D. Papaleo, José Fernando Maya-Vetencourt 외

Graphene conductive properties have been long exploited in the field of organic photovoltaics and optoelectronics by the scientific community worldwide. We engineered and characterized a hybrid biointerface in which grap…