Hybrid Memoised Wake-Sleep: Approximate Inference at the Discrete-Continuous Interface

Authors: Tuan Anh Le, Katherine M. Collins, Luke Hewitt, Kevin Ellis, Siddharth N, Samuel Gershman, Joshua B. Tenenbaum

ICLR 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate HMWS in the GP-kernel learning and 3D scene understanding domains, and show that it outperforms current state-of-the-art inference methods.
Researcher Affiliation Academia 1MIT, 2Cornell University, 3University of Edinburgh & The Alan Turing Institute, 4Harvard University
Pseudocode Yes Algorithm 1 Hybrid Memoised Wake-Sleep (a single learning iteration)
Open Source Code No No explicit statement or link providing concrete access to source code for the methodology was found.
Open Datasets No No concrete access information (specific link, DOI, repository name, formal citation with authors/year, or reference to established benchmark datasets) for a publicly available or open dataset was found.
Dataset Splits No No specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology) needed to reproduce the data partitioning was found.
Hardware Specification No No specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments were found.
Software Dependencies No No specific ancillary software details (e.g., library or solver names with version numbers like Python 3.8, CPLEX 12.4) needed to replicate the experiment were found.
Experiment Setup Yes For training, we use the Adam optimizer with default hyperparameters. and Fixed parameters: Vocabulary V = {WN, SE, PER1, . . . , PER4, C, , +, (, )} Vocabulary size vocabulary_size = |V| = 11 Kernel parameters K = {σ2 WN, (σ2 SE, ℓ2 SE), (σ2 PER1, p PER1, ℓ2 PER1), . . . , (σ2 PER4, p PER4, ℓ2 PER4), σ2 C} Hidden size hidden_size = 128 Observation embedding size obs_embedding_size = 128 and Number of allowed primitives to learn P = 5 Maximum number of blocks per cell Bmax = 3 Number of cells in the x-plane = number of cells in the z-plane = N = 2 Image resolution I = [3, 128, 128] Observation embedding size obs_embedding_size = 676