Incrementally Grounding Expressions for Spatial Relations between Objects

Authors: Tiago Mota, Mohan Sridharan

IJCAI 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The architecture is evaluated on a benchmark dataset of tabletop images and on complex simulated scenes of furniture. ... We evaluate these capabilities on a benchmark dataset of tabletop objects and complex, simulated scenes of furniture. ... For experimental evaluation, we used the Table Object Scene Database (TOSD) and simulated scenes. ... Experiments tested two hypotheses:
Researcher Affiliation Academia Tiago Mota1 and Mohan Sridharan2 1 The University of Auckland, NZ 2 University of Birmingham, UK
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks. It describes the architecture and methods in prose and with diagrams.
Open Source Code No The paper does not provide any specific links or statements about the availability of the source code for the methodology described.
Open Datasets Yes For experimental evaluation, we used the Table Object Scene Database (TOSD)3 and simulated scenes. ... 3https://repo.acin.tuwien.ac.at/tmp/permanent/TOSD.zip
Dataset Splits Yes Pairs of objects extracted from the training set of the TOSD were randomly divided into 10 subsets. Seven pairs of objects from each subset were used to train the MSR-based grounding with human feedback. ... A MSR-based grounding was acquired using QSRbased labels for four out of the five subsets ( 2000 pairs) in each run.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., CPU, GPU models, memory, or cloud instances) used to run the experiments.
Software Dependencies No The paper mentions using "SPARC system [Balai et al., 2013]" and "Euclidean cluster extraction segmentation algorithm [Rusu, 2010]" and "Bullet physics library" but does not specify version numbers for these software components.
Experiment Setup Yes The first set of experiments was designed as follows, with the results summarized in Table 1: 1. Pairs of objects extracted from the training set of the TOSD were randomly divided into 10 subsets. 2. Seven pairs of objects from each subset were used to train the MSR-based grounding with human feedback. ... The second set of experiments was designed as follows, with the results summarized in Table 2: 1. Pairs of objects extracted from the training set of the TOSD were randomly divided into five subsets. 2. A MSR-based grounding was acquired using QSRbased labels for four out of the five subsets ( 2000 pairs) in each run.