Visual Memory for Robust Path Following
Authors: Ashish Kumar, Saurabh Gupta, David Fouhey, Sergey Levine, Jitendra Malik
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our experiments show that our approach outperforms classical approaches and other learning based baselines. |
| Researcher Affiliation | Academia | Ashish Kumar Saurabh Gupta David Fouhey Sergey Levine Jitendra Malik University of California, Berkeley ashish_kumar@berkeley.edu, {sgupta, dfouhey, svlevine, malik}@eecs.berkeley.edu |
| Pseudocode | No | The paper describes the proposed architecture and its components textually and with diagrams, but does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | Project website with videos: https://ashishkumar1993.github.io/rpf/ |
| Open Datasets | Yes | The first simulator is based on real world scans from Stanford Building Parser Dataset [2] (SBPD) and the Matterport 3D Dataset [7](MP3D). The second simulation environment is based on SUNCG [39]. |
| Dataset Splits | Yes | We use splits that ensure that the testing environment comes from an entirely different building than the training or validation environments. Once again, splits ensure no overlap between training and testing environments. |
| Hardware Specification | No | The paper does not provide specific details on the hardware used for experiments, such as GPU or CPU models. It only mentions network architectures like Convolutional Networks and GRUs. |
| Software Dependencies | No | The paper mentions 'Adam optimizer' and 'COLMAP package', but it does not specify version numbers for these or any other software dependencies required for reproducibility. |
| Experiment Setup | Yes | We train the entire network from scratch in an end-to-end manner using Adam optimizer for 120000 iterations, where each episode is 40 steps long. φ is a 5 layer Convolutional Network with [32, 64, 128, 256, 512] filters respectively. |