Scalable methods to integrate task knowledge with the Three-Weight Algorithm for hybrid cognitive processing via optimization

Derbinsky, Nate; Bento, José; Yedidia, Jonathan S.

doi:10.1016/j.bica.2014.03.007

Cited by 4 publications

(7 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…on the fact that ADMM (i) allows us to easily parallelize BIG-BOSS, (ii) has good empirical performance for several nonsmooth, nonconvex problems [4,11,12,20,29,45], and (iii) has global optimality guarantees (under mild convexity assumptions) and a convergence rate that (under proper tuning) equals the convergence rate of the fastest possible first-order method [16], although itself not a first-order method.…”

Section: Solution Procedures Using Admmmentioning

confidence: 99%

“…. , f (12) be defined analogously. For the special case of constraints (9), we define one function f (i) (9) for each condition i in (9).…”

Section: Solution Procedures Using Admmmentioning

confidence: 99%

“…For the special case of constraints (9), we define one function f (i) (9) for each condition i in (9). To explain how we solve problem (6)-(12) using ADMM, it is convenient to perform two simple rerepresentations of (6)- (12) using the functions f i .…”

Section: Solution Procedures Using Admmmentioning

confidence: 99%

“…There are 3 + k function-nodes and variablenodes in total. We interpret ADMM as an iterative scheme that operates on iterates that live on the edges/nodes of the factor-graph in Figure 1, similar to the approaches in [11,12,20]. The function-nodes have the following labels: quadratic (QP), sparse (SP), bi-linear (BI (1) , .…”

Section: Solution Procedures Using Admmmentioning

confidence: 99%

“…Observe that the constraint f (11) + f (12) < +∞ is equivalent to 3k constraints of the type e (i, j) ≤ w (i, j) ≤ q (i, j) for some constant vectors e (i, j) , q (i, j) ∈ R n : one for each condition i and each of the variables v (i) (for index j = 1), µ (i) ( for index j = 3), and η (i) (for index j = 4). Hence, computing Prox DB ({n (i, j) }) for some input {n (i, j) } ∈ R 3kn amounts to solving 3k independent problems min…”

Section: Proximal Operator For Node Dbmentioning

confidence: 99%

See 4 more Smart Citations

Estimating Cellular Goals from High-Dimensional Biological Data

Yang

Saunders

Lachance

et al. 2019

Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery &Amp; Data Mining

Self Cite

View full text Add to dashboard Cite

Optimization-based models have been used to predict cellular behavior for over 25 years. The constraints in these models are derived from genome annotations, measured macro-molecular composition of cells, and by measuring the cell's growth rate and metabolism in different conditions. The cellular goal (the optimization problem that the cell is trying to solve) can be challenging to derive experimentally for many organisms, including human or mammalian cells, which have complex metabolic capabilities and are not well understood. Existing approaches to learning goals from data include (a) estimating a linear objective function, or (b) estimating linear constraints that model complex biochemical reactions and constrain the cell's operation. The latter approach is important because often the known/observed biochemical reactions are not enough to explain observations, and hence there is a need to extend automatically the model complexity by learning new chemical reactions. However, this leads to nonconvex optimization problems, and existing tools cannot scale to realistically large metabolic models. Hence, constraint estimation is still used sparingly despite its benefits for modeling cell metabolism, which is important for developing novel antimicrobials against pathogens, discovering cancer drug targets, and producing value-added chemicals. Here, we develop the first approach to estimating constraint reactions from data that can scale to realistically large metabolic models. Previous tools have been used on problems having less than 75 biochemical reactions and 60 metabolites, which limits real-life-size applications. We perform extensive experiments using 75 large-scale metabolic network models for different organisms (including bacteria, yeasts, and mammals) and show that our algorithm can recover cellular constraint reactions. The recovered constraints enable accurate prediction of metabolic states in hundreds of growth environments not seen in training data, and we recover useful cellular goals even when some measurements are missing. * Both authors contributed equally to this project and are corresponding authors.

show abstract

Section: Solution Procedures Using Admmmentioning

confidence: 99%

“…. , f (12) be defined analogously. For the special case of constraints (9), we define one function f (i) (9) for each condition i in (9).…”

Section: Solution Procedures Using Admmmentioning

confidence: 99%

Section: Solution Procedures Using Admmmentioning

confidence: 99%

Section: Solution Procedures Using Admmmentioning

confidence: 99%

Section: Proximal Operator For Node Dbmentioning

confidence: 99%

See 3 more Smart Citations

Estimating Cellular Goals from High-Dimensional Biological Data

Yang

Saunders

Lachance

et al. 2019

Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery &Amp; Data Mining

Self Cite

View full text Add to dashboard Cite

show abstract

Expression Graphs

Demski

2015

Artificial General Intelligence

View full text Add to dashboard Cite

Testing Fine-Grained Parallelism for the ADMM on a Factor-Graph

Hao

Oghbaee

Rostami

et al. 2016

2016 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)

Self Cite

View full text Add to dashboard Cite

Abstract-There is an ongoing effort to develop tools that apply distributed computational resources to tackle large problems or reduce the time to solve them. In this context, the Alternating Direction Method of Multipliers (ADMM) arises as a method that can exploit distributed resources like the dual ascent method and has the robustness and improved convergence of the augmented Lagrangian method. Traditional approaches to accelerate the ADMM using multiple cores are problem-specific and often require multi-core programming. By contrast, we propose a problem-independent scheme of accelerating the ADMM that does not require the user to write any parallel code. We show that this scheme, an interpretation of the ADMM as a message-passing algorithm on a factor-graph, can automatically exploit fine-grained parallelism both in GPUs and shared-memory multi-core computers and achieves significant speedup in such diverse application domains as combinatorial optimization, machine learning, and optimal control. Specifically, we obtain 10-18x speedup using a GPU, and 5-9x using multiple CPU cores, over a serial, optimized C-version of the ADMM, which is similar to the typical speedup reported for existing GPU-accelerated libraries, including cuFFT (19x), cuBLAS (17x), and cuRAND (8x).

show abstract

Scalable methods to integrate task knowledge with the Three-Weight Algorithm for hybrid cognitive processing via optimization

Cited by 4 publications

References 20 publications

Estimating Cellular Goals from High-Dimensional Biological Data

Estimating Cellular Goals from High-Dimensional Biological Data

Expression Graphs

Testing Fine-Grained Parallelism for the ADMM on a Factor-Graph

Contact Info

Product

Resources

About