Lifted Inference Rules With Constraints

Authors: Happy Mittal, Anuj Mahajan, Vibhav G. Gogate, Parag Singla

NeurIPS 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments on benchmark MLNs for exact and sampling based inference demonstrate the effectiveness of our approach over several other existing techniques.
Researcher Affiliation Academia Happy Mittal, Anuj Mahajan Dept. of Comp. Sci. & Engg. I.I.T. Delhi, Hauz Khas New Delhi, 110016, India Vibhav Gogate Dept. of Comp. Sci. Univ. of Texas Dallas Richardson, TX 75080, USA Parag Singla Dept. of Comp. Sci. & Engg. I.I.T. Delhi, Hauz Khas New Delhi, 110016, India
Pseudocode No The paper refers to "Algorithm 2" and "our partitioning algorithm" and states "We refer the reader to the supplement section for a detailed description of our partitioning algorithm." but no pseudocode or algorithm block is present in the provided text.
Open Source Code No The paper mentions "Both our systems and GCFOVE are implemented in Java. PTP is implemented in C++." and provides links to "Alchemy-2:code.google.com/p-alchemy-2,GCFOVE: https:dtai.cs.kuleuven.be/software/gcfove". However, these links are for the systems they compared against, not for their own "Set In Eq" approach. No explicit statement or link for their own source code is provided.
Open Datasets No The paper mentions using "benchmark MLN domains" such as "Friends & Alchemy Smokes", "Web KB", and "IMDB", and provides a table with their details (Table 2). However, it does not provide concrete access information (e.g., links, DOIs, or specific repositories) for these datasets.
Dataset Splits No The paper does not explicitly provide information on training/validation/test dataset splits (e.g., percentages or sample counts) for its experiments.
Hardware Specification Yes The experiments on all the datasets except Web KB were carried on a machine with 2.20GHz Intel Core i3 CPU and 4GB RAM. Web KB is a much larger dataset and we ran the experiments on 2.20 GHz Xeon(R) E5-2660 v2 server with 10 cores and 128 GB RAM.
Software Dependencies No The paper states "Both our systems and GCFOVE are implemented in Java. PTP is implemented in C++." but does not provide specific version numbers for Java or C++, nor does it list any specific libraries with version numbers.
Experiment Setup Yes For approximate inference in both Normal and Set In Eq, we used the unbiased importance sampling scheme as described by Gogate & Domingos [11]. We collected a total of 1000 samples for each estimate and averaged the Z values.