Focus of Attention Improves Information Transfer in Visual Features
Authors: Matteo Tiezzi, Stefano Melacci, Alessandro Betti, Marco Maggini, Marco Gori
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We provide experimental results to support the theory by showing that the spatio-temporal filtering induced by the focus of attention allows the system to globally transfer more information from the input stream over the focused areas and, in some contexts, over the whole frames with respect to the unfiltered case that yields uniform probability distributions. |
| Researcher Affiliation | Academia | 1DIISM, University of Siena, Siena, Italy 2Maasai, Universitè Côte d Azur, Nice, France |
| Pseudocode | No | The paper does not include any sections or figures explicitly labeled as 'Pseudocode' or 'Algorithm', nor are there any code-like structured steps. |
| Open Source Code | Yes | A Py Torch-based implementation can be downloaded as supplementary material. |
| Open Datasets | No | The paper mentions using 'SPARSEMNIST', 'CARPARK', and 'CALL' video streams but does not provide specific links, DOIs, repository names, or formal citations for public access to these datasets. |
| Dataset Splits | No | The paper specifies a split for learning (100k frames) and testing (5k frames) but does not explicitly mention a distinct validation set or its split details. |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., GPU models, CPU types, memory) used to run the experiments. |
| Software Dependencies | No | The paper mentions a 'Py Torch-based implementation' but does not provide specific version numbers for PyTorch or any other software dependencies. |
| Experiment Setup | Yes | After a first experimentation in which we qualitatively observed the behaviour of the 2nd order laws, we set α = 0.01, β = 0.1, k = 10^-8. For each model we considered multiple weighing schemes of the parameters λc {10, 100, 200, 1000}, λe {20, 200, 400, 2000, 4000}, λs {10, 100, 1000}, { 0.01, 0.05, 0.07}, selecting the ones that returned the largest MI during the learning stage. |