Fine-tuned 7B LLMs generating unified diffs for neural architecture refinement achieve 66-75% valid rates and 64-66% mean first-epoch accuracy, outperforming full-generation baselines by large margins while cutting output length by 75-85%.
Preparation of Fractal-Inspired Computational Architectures for Automated Neural Design Exploration
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
abstract
This paper proposes FractalNet, a framework based on fractal design principles that automatically generates and evaluates convolutional neural network (CNN) architectures using recursive template patterns. Rather than relying on computationally expensive Neural Architecture Search (NAS) methods, the framework explores a structured architecture space defined by recursive fractal templates that systematically vary key parameters such as fractal depth, column width, and layer configurations. The framework consists of three core components: a generator that produces candidate architectures via controlled permutations of convolutional, normalization, activation, and dropout layers; a fractal template module that enforces recursive multi-path structural patterns; and a runner module that manages model training, evaluation, and logging. Using this system, over 1,200 distinct CNN architectures were automatically generated and evaluated on the CIFAR-10 image classification benchmark. Training was performed in PyTorch using stochastic gradient descent with Automatic Mixed Precision (AMP) and gradient checkpointing to reduce computational overhead. Experimental results demonstrate that fractal-based architectures exhibit stable training dynamics and achieve competitive performance, with an average validation accuracy of 60-70% and a peak accuracy of 80.18% after only five training epochs. These findings suggest that recursive fractal structures provide an effective means of balancing network depth and width while supporting large-scale automated architecture exploration. The proposed framework offers a resource-efficient and interpretable approach to systematic neural architecture experimentation.
representative citing papers
Closed-loop LLM search with AST-generated examples discovers non-standard channel widths that improve vision model performance over initial architectures on CIFAR-100.
Three-example few-shot prompting optimizes LLM-generated vision architectures while a whitespace-normalized hash provides 100x faster duplicate detection than AST parsing across seven benchmarks.
LEMUR 2 releases a multi-generator, multi-task neural-architecture corpus with real-device latency metadata intended as fuel for LLM-driven AutoML.
Automated search of 4463 heterogeneous 4-expert MoE models found enumeration bias anchoring the space to AirNet and ranked ShuffleNet/MobileNetV3 as top performers.
Empirical grid search over 18 loss-optimizer pairs on 33 LEMUR architectures shows cross-entropy with Adam/AdamW is most robust while NGL and SGD-based pairings vary sharply by model family.
citing papers explorer
-
Delta-Based Neural Architecture Search: LLM Fine-Tuning via Code Diffs
Fine-tuned 7B LLMs generating unified diffs for neural architecture refinement achieve 66-75% valid rates and 64-66% mean first-epoch accuracy, outperforming full-generation baselines by large margins while cutting output length by 75-85%.
-
Closed-Loop LLM Discovery of Non-Standard Channel Priors in Vision Models
Closed-loop LLM search with AST-generated examples discovers non-standard channel widths that improve vision model performance over initial architectures on CIFAR-100.
-
Enhancing LLM-Based Neural Network Generation: Few-Shot Prompting and Efficient Validation for Automated Architecture Design
Three-example few-shot prompting optimizes LLM-generated vision architectures while a whitespace-normalized hash provides 100x faster duplicate detection than AST parsing across seven benchmarks.
-
LEMUR 2: Unlocking Neural Network Diversity for AI
LEMUR 2 releases a multi-generator, multi-task neural-architecture corpus with real-device latency metadata intended as fuel for LLM-driven AutoML.
-
Systematic Exploration of 4-Expert Heterogeneous Mixture-of-Experts via Automated Pipeline Search
Automated search of 4463 heterogeneous 4-expert MoE models found enumeration bias anchoring the space to AirNet and ranked ShuffleNet/MobileNetV3 as top performers.
-
Towards Robust Training in NNGPT AutoML Pipeline: A Loss-Optimizer Pairing Selection Study
Empirical grid search over 18 loss-optimizer pairs on 33 LEMUR architectures shows cross-entropy with Adam/AdamW is most robust while NGL and SGD-based pairings vary sharply by model family.