Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Improving Variational Methods via Pairwise Linear Response Identities
Authors: Jack Raymond, Federico Ricci-Tersenghi
JMLR 2017 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In Section 4, we compare the performance of constrained approximations against exact results on some standard models, demonstrating a significant advantage in many cases. |
| Researcher Affiliation | Academia | Jack Raymond EMAIL Dipartimento di Fisica, La Sapienza University of Rome Piazzale Aldo Moro 5 Rome, Italy Federico Ricci-Tersenghi EMAIL Dipartimento di Fisica, INFN Sezione di Roma1 and CNR Nanotec, La Sapienza University of Rome Piazzale Aldo Moro 5 Rome, Italy |
| Pseudocode | Yes | Appendix B.5 Pseudocode for Determining the Fixed Point of q Given λ, Algorithm 1 λ compatible Heskes-Albers-Kappen algorithm |
| Open Source Code | No | lib DAI is a code repository that has collected some of the methods together (see Mooij, 2010), we developed our methods based on this library, in particular, the implementation of Heskes, Albers, and Kappen (2003). The paper discusses using a third-party library but does not provide concrete access to the authors' own source code for the methodology described. |
| Open Datasets | Yes | The alarm net is a pedagogical example of a graphical model that has been studied in the context of loop correction algorithms and is available in lib DAI repository (see Mooij et al., 2007; Mooij, 2010). |
| Dataset Splits | No | The paper evaluates inference methods on "toy model frameworks" and "random 3-regular graphs" without involving machine learning style training/test/validation splits. No specific dataset split information is provided. |
| Hardware Specification | No | The paper does not provide specific hardware details (exact GPU/CPU models, processor types, memory amounts, or detailed computer specifications) used for running its experiments. |
| Software Dependencies | No | lib DAI: A free and open source c++ library for discrete approximate inference in graphical models. The paper mentions "lib DAI" and that they developed their methods based on this library, but it does not specify a version number for lib DAI or any other software dependencies. |
| Experiment Setup | Yes | In Section 4.1 we study a fully connected ferromagnetic model (that is a model with a positive coupling between any pair of variables) with symmetry broken (that is with a nonzero mean value for each variable)... N = 10 and h = 1... For the Wainwright-Jordan set-up... Fields hi are independent and identically distributed (iid) samples from [ 0.25, 0.25] and couplings Jij are sampled i.i.d on [ 1, 1]... We considered 20 instances for L = 4 and L = 7... annealing procedure is employed... damping within the 3-cycle of Figure 3 can be necessary for convergence. In practice, we replace λt+1 = dλt + (1 d)λ , where λ is the cavity approximation. As 1/T was increased (or decreased) at a constant rate we increased d whenever the solution failed to converge... Initial conditions are chosen as λ = 0 |