FiLM: Visual Reasoning with a General Conditioning Layer

Abstract : We introduce a general-purpose conditioning method for neu-ral networks called FiLM: Feature-wise Linear Modulation. FiLM layers influence neural network computation via a simple , feature-wise affine transformation based on conditioning information. We show that FiLM layers are highly effective for visual reasoning — answering image-related questions which require a multi-step, high-level process — a task which has proven difficult for standard deep learning methods that do not explicitly model reasoning. Specifically, we show on visual reasoning tasks that FiLM layers 1) halve state-of-the-art error for the CLEVR benchmark, 2) modulate features in a coherent manner, 3) are robust to ablations and architectural modifications, and 4) generalize well to challenging, new data from few examples or even zero-shot.
Complete list of metadatas

Cited literature [32 references]  Display  Hide  Download

https://hal.inria.fr/hal-01648685
Contributor : Florian Strub <>
Submitted on : Tuesday, November 28, 2017 - 3:13:55 AM
Last modification on : Friday, March 22, 2019 - 1:36:13 AM

Links full text

Identifiers

  • HAL Id : hal-01648685, version 1
  • ARXIV : 1707.03017

Citation

Ethan Perez, Florian Strub, Harm de Vries, Vincent Dumoulin, Aaron Courville. FiLM: Visual Reasoning with a General Conditioning Layer. AAAI Conference on Artificial Intelligence, Feb 2018, New Orleans, United States. ⟨hal-01648685⟩

Share

Metrics

Record views

336