Active-set Methods for Submodular Minimization Problems

Authors: K. S. Sesh Kumar, Francis Bach

JMLR 2017 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In the experiments, we compare the performance of the proposed algorithms with state-of-the-art algorithms, showing that it reduces the calls to SFM oracles. ... Section 5. Experiments: In this section, we show the results of the algorithms proposed on various problems.
Researcher Affiliation Academia K. S. Sesh Kumar, Imperial College London, Department of Computing. Francis Bach, INRIA Sierra Project-team D epartement d Informatique de l Ecole Normale Sup erieure (UMR CNRS/ENS/INRIA).
Pseudocode Yes 3.4 Active-set Algorithm: Data: Submodular function F with SFM oracle, u Rn, ordered partition A Result: primal optimal: w Rn and dual optimal: s B(F) ... 4.3 Active-set Algorithm for Decomposable Functions: Data: Submodular functions F1 and F2 with SFM oracles, u Rn, ordered partitions A1, A2 Result: primal optimal: w Rn and dual optimal: s1 B(F1), s2 B(F2) ... 4.4.2 PRIMAL ACTIVE-SET METHOD: ... Primal active-set algorithm.
Open Source Code No The paper does not provide a specific link to source code, nor does it explicitly state that the code for their methodology is open-source or available in supplementary materials.
Open Datasets Yes In our experiments, we consider the 3D volumetric data set of the Stanford bunny (max) of size 102 100 79. ... Maxflow dataset online. http://vision.csd.uwo.ca/maxflow-data.
Dataset Splits No The paper describes the datasets used (various images and the Stanford bunny), but does not provide any explicit information regarding training, testing, or validation splits for these datasets.
Hardware Specification Yes 3. 20 core, Intel(R) Xeon(R) CPU E5-2670 v2 @ 2.50GHz with 100Gigabytes of memory. We only use up to 16 cores of the machine to ensure accurate timings.
Software Dependencies No The paper mentions the use of 'Maxflow' as an SFM oracle and refers to several algorithms by name, but does not provide specific version numbers for any software dependencies, libraries, or frameworks used in their implementation.
Experiment Setup Yes The unweighted total variation is f(w) = λ X i j |wi wj|, where λ is a regularizing constant for solving the total variation problem in Eq. 3. ... The unary potentials of each pixel is calculated using the Gaussian mixture model of the color features. The edge weight a(i, j) = exp( yi yj 2), where yi denotes the RGB values of the pixel i. ... We use 200 and 500 regions. ... Therefore, we introduce a parameter α to approximately solve the Dykstra step such that s1 + s2 u + w 1 α(ϵ1 + ϵ2).