# Contextual Feedback Loops: Amplifying Deep Reasoning with Iterative Top-Down Feedback

arXiv: https://arxiv.org/abs/2412.17737 · HTML: https://mlsanity.com/paper/2412.17737/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 22 references that are in the map.
- First posted on arXiv 23 Dec 2024.

## Builds on (most cited first, 12 of 12)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [Compressive Transformers for Long-Range Sequence Modelling](https://mlsanity.com/paper/1911.05507/) (2019)
- [Long Range Arena: A Benchmark for Efficient Transformers](https://mlsanity.com/paper/2011.04006/) (2020)
- [Dynamic Routing Between Capsules](https://mlsanity.com/paper/1710.09829/) (2017)
- [Deep Equilibrium Models](https://mlsanity.com/paper/1909.01377/) (2019)
- [Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning](https://mlsanity.com/paper/1605.08104/) (2016)
- [DRAW: A Recurrent Neural Network For Image Generation](https://mlsanity.com/paper/1502.04623/) (2015)
- [Neural Networks with Recurrent Generative Feedback](https://mlsanity.com/paper/2007.09200/) (2020)
- [Predify: Augmenting deep neural networks with brain-inspired predictive coding dynamics](https://mlsanity.com/paper/2106.02749/) (2021)
- [Model-based Iterative Restoration for Binary Document Image Compression with Dictionary Learning](https://mlsanity.com/paper/1704.07019/) (2017)

## Built on by (most cited first, 1 of 1)

- [Iterate to Accelerate: A Unified Framework for Iterative Reasoning and Feedback Convergence](https://mlsanity.com/paper/2502.03787/) (2025)

## Abstract (arXiv)

Conventional deep networks rely on one-way backpropagation that overlooks reconciling high-level predictions with lower-level representations. We propose Contextual Feedback Loops (CFLs), a lightweight mechanism that re-injects top-down context into earlier layers for iterative refinement. Concretely, CFLs map the network's prediction to a compact context vector, which is fused back into each layer via gating adapters. Unrolled over multiple feedback steps, CFLs unify feed-forward and feedback-driven inference, letting top-level outputs continually refine lower-level features. Despite minimal overhead, CFLs yield consistent gains on tasks including CIFAR-10, ImageNet-1k, SpeechCommands, and GLUE SST-2. Moreover, by a Banach Fixed Point argument under mild Lipschitz conditions, these updates converge stably. Overall, CFLs show that even modest top-down feedback can substantially improve deep models, aligning with cognitive theories of iterative perception.