PROBLP: A framework for Iow-precision probabilistic inference

Bayesian reasoning is a powerful mechanism for probabilistic inference in smart edge-devices. During such inferences, a low-precision arithmetic representation can enable improved energy efficiency. However, its impact on inference accuracy is not yet understood. Furthermore, general-purpose hardware does not natively support low-precision representation. To address this, we propose PROBLP, a framework that automates the analysis and design of low-precision probabilistic inference hardware. It automatically chooses an appropriate energy-efficient representation based on worst-case errorbounds and hardware energy-models. It generates custom hardware for the resulting inference network exploiting parallelism, pipelining and low-precision operation. The framework is validated on several embedded-sensing benchmarks.

[1]  Marian Verhelst,et al.  Dynamic Sensor-Frontend Tuning for Resource Efficient Embedded Classification , 2018, IEEE Journal on Emerging and Selected Topics in Circuits and Systems.

[2]  Adnan Darwiche,et al.  When do Numbers Really Matter? , 2001, UAI.

[3]  Nicholas J. Higham,et al.  INVERSE PROBLEMS NEWSLETTER , 1991 .

[4]  Adnan Darwiche,et al.  Sensitivity Analysis in Bayesian Networks: From Single to Multiple Parameters , 2004, UAI.

[5]  Marian Verhelst,et al.  Towards Resource-Efficient Classifiers for Always-On Monitoring , 2018, ECML/PKDD.

[6]  Sebastian Tschiatschek,et al.  Bayesian Network Classifiers with Reduced Precision Parameters , 2012, ECML/PKDD.

[7]  Guy Van den Broeck,et al.  Learning Logistic Circuits , 2019, AAAI.

[8]  Gregory F. Cooper,et al.  The ALARM Monitoring System: A Case Study with two Probabilistic Inference Techniques for Belief Networks , 1989, AIME.

[9]  Davide Anguita,et al.  A Public Domain Dataset for Human Activity Recognition using Smartphones , 2013, ESANN.

[10]  Adnan Darwiche,et al.  On probabilistic inference by weighted model counting , 2008, Artif. Intell..

[11]  Luc De Raedt,et al.  DeepProbLog: Neural Probabilistic Logic Programming , 2018, BNAIC/BENELEARN.

[12]  Petia Radeva,et al.  Personalization and user verification in wearable systems using biometric walking patterns , 2011, Personal and Ubiquitous Computing.

[13]  V. S. Costa,et al.  Theory and Practice of Logic Programming , 2010 .

[14]  David D. Wentzloff,et al.  Hardware Accelerator for Probabilistic Inference in 65-nm CMOS , 2016, IEEE Transactions on Very Large Scale Integration (VLSI) Systems.

[15]  Luc De Raedt,et al.  Inference and learning in probabilistic logic programs using weighted Boolean formulas , 2013, Theory and Practice of Logic Programming.

[16]  Daniela Micucci,et al.  UniMiB SHAR: a new dataset for human activity recognition using acceleration data from smartphones , 2016, ArXiv.

[17]  Judea Pearl,et al.  Chapter 2 – BAYESIAN INFERENCE , 1988 .

[18]  Bhaskara Marthi,et al.  A generative vision model that trains with high data efficiency and breaks text-based CAPTCHAs , 2017, Science.

[19]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.

[20]  Luc De Raedt,et al.  Inference in Probabilistic Logic Programs using Weighted CNF's , 2011, UAI.

[21]  Jean-Philippe Diguet,et al.  FPGA implementation of Bayesian network inference for an embedded diagnosis , 2015, 2015 IEEE Conference on Prognostics and Health Management (PHM).