Preprint

  • Lysy, M., Asgharian M. and Partovi Nia, V. (2020) A Convergence Diagnostic for Bayesian Clustering , submitted, under review.
  • Bayesian clustering lacks a concrete convergence criterion. Here we generalize the chi-square goodness of fit statistic to apply it for Markov chains run on groupings.
  • Murua, A., Ramakrishnan, R.K., Li, X., Yang, R. H., and Partovi Nia, V. (2020) Tensor Train Decompositions on Recurrent Networks submitted, under review.
  • Here we show Matrix Product State tensor trains are easier to train, and more appropriate for compression and inference speed up compared to Matrix Product Operator tensor trains.
  • Partovi Nia, V., Li, X., Asgharian, M., Hu, S., Chen, Z., and Geng, Y. (2020) Clustering Causal Additive Noise Models submitted, under review.
  • We show how causal direction test statistics loses its accuracy in the presence of heterogenous data. We provide a remedy that works with unknown number of clusters.

  • Sari, E., Belbahri, M., and Partovi Nia, V. (2019) How Does Batch Normalization Help Binary Training?, submitted, under review.
  • Here we study how to initialize binary quantized neural networks, while common full-precision initializations fail.

  • Tilouche, S., Partovi Nia, V., and Bassetto, S. (2019) Parallel Coordinate Order for High-Dimensional Data, submitted, under review.
  • We show how to visualize high-dimensional data properly.

  • Belbahri, M., Murua, A., Gandouet, O., and Partovi Nia, V. (2019) Uplift Regression: The R Package tools4uplift, submitted, under review.
  • We show how to use our R package to fit uplift regression models.

    Published

  • Sari, X., and Partovi Nia, V. (2020) Batch Normalization in Quantized Networks, accepted, to appear in Edge Intelligence 2020 Workshop Proceedings.
  • Most studies on BatchNorm are focused on full-precision networks, and there is little research in understanding BatchNorm affect in quantized training which we address here. We extend JCVIS in ternary case work and study binary and ternary within the same theoretical framework.
  • Li, X., and Partovi Nia, V. (2020) Random Bias Initialization Improves Quantized Training, accepted, to appear in Edge Intelligence 2020 Workshop Proceedings.
  • Training quantized networks is much longer than full precision. We study the geometry of back-propagation of quantized network and argue that random bias initialization speeds up their training.
  • Zolnouri, M. , Li, X., and Partovi Nia, V. (2020) Importance of Data Loading Pipeline in Training Deep Neural Networks, accepted, to appear in Edge Intelligence 2020 Workshop Proceedings.
  • We show how to speed up training by leveraging on data loaders.
  • Ramakrishnan, R. K., Sari, E., and Partovi Nia, V. (2020) Differentiable Mask Pruning for Neural Networks, accepted, to appear in Computer and Robot Vision (CRV).
  • We show how to train deep networks while pruning unnecessary parameters, while matching forward and backward pass through the differentiable mask.
  • Farhadi, F., Partovi Nia, V., and Lodi, A. (2020) Activation Adaptation in Neural Networks, Proceedings of the 9th International Conference on Pattern Recognition Applications and Methods (ICPRAM) Vol 1, 249-257.
  • Learning the activation function improves prediction accuracy in neural networks.
  • Sari, E. and Partovi Nia, V. (2020) Understanding BatchNorm in Ternary Training, Journal of Computational Vision and Imaging Systems, Vol 5 No 1, 2.
  • It is well-known that ternary training without BatchNorm layer is infeasible. Here we study what is the role of the BatchNorm layer in ternary quantized network.
  • Courville, V. and Partovi Nia, V. (2020) Deep Learning Inference Frameworks for ARM CPU, Journal of Computational Vision and Imaging Systems, Vol 5 No 1, 3.
  • This is an expository work that reviews existing frameworks for implementation of deep models on ARM CPU.
  • Darabi, S., Belbahri, M., Courbariaux, M., Partovi Nia, V. (2019) Regularized Binary Network Training , NeurIPS Workshop on Energy Efficient Machine Learning and Cognitive Computing.
  • We show deep neural network quantization using our regularizer, and our sign-swish activation beats state of the art compressed networks.
  • Belbahri, M., Sari, E., Darabi, S., and Partovi Nia, V. (2019) Foothill: A Quasiconvex Regularization Function , Lecture Notes in Computer Science 11663, International Conference on Image Analysis and Recognition (ICIAR) Part II, 3-14.
  • The foothill function can be regarded as a regularizer, as a quantizer, as a loss function, or even defines a class of flexible distributions.
  • Ramakrishnan, R. K., Jui, S. and Partovi Nia, V., (2019) Deep Demosaicing for Edge Implementation, Lecture Notes in Computer Science 11663, International Conference on Image Analysis and Recognition (ICIAR) Part I, 257-286.
  • We demonstrate a simple search beats the state of the art deep learning architectures for demosicing problem.
  • Vahdat, A., Belbahri, M., and Partovi Nia, V., (2019) Active Learning for High-Dimensional Binary Features, submitted, to appear in Conference on Network and Service Management (CNSM).
  • We demonstrate how to gather data efficiently for optical fiber amplifiers.
  • Davtalab-Olyaie, M., Asgharian, M., and Partovi Nia, V. (2019) Stochastic ranking and dominance in DEA, International Journal of Production Economics, published online.
  • Pareto frontiers in Data Envelopment Analysis (DEA) are treated as deterministic. Data come from a random distributin. Here we show how to adapt the deterministic concepts to the stochastic version. Pareto frontiers in Data Envelopment Analysis (DEA) are treated as deterministic. Data come from a random distribution. Here we show how to adapt the deterministic concepts to the stochastic version.
  • Partovi Nia, V., and Belbahri, M. (2018) Binary Quantizer, Journal of Computational Vision and Imaging Systems, Vol 4 No 1.
  • Here we show how to modify the objective function in back-propagation to quantize deep neural networks into one bit with a scaling factor using a quasi convex base.
  • Hu, S., Chen, Z., Partovi Nia, V., Chan, L., Geng, Y. (2018) Causal Inference and Mechanism Clustering of a Mixture of Additive Noise Models, Advances in Neural Information Processing Systems 31, 5212-5222.
  • Here we develop clustering of functions observed with noise.
  • Ghaemi, M.S. et al. (2018) Multiomics Modeling of the Immunome, Transcriptome, Microbiome, Proteome, and Metabolome Adaptations During Human Pregnancy, Bioinformatics , Vol. 35, Issue 1, 95–103,.
  • This work shows how stacked generalization increases prediction power on pregnancy data.
  • Mirshahi, M., Partovi Nia, V., Adjengue, L. (2018) Automatic odor prediction for electronic nose, Journal of Applied Statistics, to appear.
  • This work develops a statistical pattern recognition method as a tool for odor prediction.
  • Ghaemi, M.S., Agard, B., Trepanier, M. and Partovi Nia, V. (2017) A Visual Segmentation Method for Temporal Smart Card Data, Transportmetrica A: Transport Science, Vol 13, Issue 5, 381-404.
  • This work is an intuitive visual map from a binary vector into a three-dimensional clock-like three-dimensional space to reveal the underlying temporal pattern of public transit users.
  • Mirshahi, M., Partovi Nia, V., and Adjengue, L. (2017) An Online Data Validation Algorithm for Electronic Nose, Lecture Notes in Computer Science 10163, Proceedings of International Conference on Pattern Recognition Applications and Methods (ICPRAM), 104-120.
  • This is the extended version of the ICPRAM paper with some discussions about the computational complexity of our online outlier detection algorithm.
  • Mirshahi, M., Partovi Nia, V., and Adjengue, L. (2016) Statistical Measurement Validation with Application to Electronic Nose Technology, Proceedings of the International Conference on Pattern Recognition Applications and Methods (ICPRAM), 407-414.
  • How do we predict odors? It seems easy at the first sight, since odor and chemicals are closely related, but a mixture of several chemicals changes the odor perception. Sensors data are often unreliable. Here we describe and automatic method to validate the sensors data.
  • Partovi Nia, V. and Ghannad-Rezaie, M. (2016) Agglomerative Joint Clustering of Metabolic Data with Spike at Zero: A Bayesian Perspective, Biometrical Journal, Vol 58, Issue 2, 387-396.
  • Dendrogram is a visualization tool to demonstrate evolution of data groupings. We generalized dendrogram to forestogram.
  • Homaie, A. H., Partovi Nia, V., Gamache, M. and Agard, B. (2016) Flight deck crew reserve: from data to forecasting, Engineering Applications of Artificial Intelligence, Vol 50, 106-114.
  • We develop a new version of a the random forest to predict the flight crew absence of airlines.
  • Davtalab-Olyaie, M., Roshdi, I., and Partovi Nia, V., and Asgharian, M. (2015) On Characterizing Full Dimensional Weak Facets in DEA with Variable Returns to Scale Technology, Optimization, Vol 64, Issue 11, 2455-2476.
  • Data envelopment analysis is a widely used technique for computing the relative efficiency over decision making units. It turns out that the frontier of the envelop is composed of two sort of facets, strong facts and weak facets. We developed a theory for characterizing the weak facets of the envelop and provided a mixed-integer programming that computes all of them.
  • Kijko, G., Jolliet, O., Partovi Nia, V., Doudrich G., and Margni M. (2015) Impact of occupational exposure to chemicals in life cycle assessment: A novel characterization model based on measured concentrations and labour hours, Environmental Science and Technology, Vol 49, Issue 14, 8741-8750.
  • We show how uncertainty can be quantified using Monte Carlo method in life cycle data analysis.
  • Partovi Nia, V. and Davison, A. C. (2015) A Simple Model-Based Approach to Variable Selection in Classification and Clustering, Canadian Journal of Statistics, Vol 43, Issue 2, 157-175.
  • We show that a simple Bayesian variable selection adapted for classification and clustering is comparable with many sophisticated exisiting variable selection methods.
  • Ghaemi, M. S., Agard, B., Partovi Nia, V., and Trepanier, M. (2015) Challenges in Spatial-Temporal Data Analysis Targeting Public Transport, 15th IFAC Symposium on Information Control Problems in Manufacturing (INCOM), Vol 48, Issue 3, 442-447.
  • We discuss complex challenges in public transport data mining and resolve some of the issues with simple tricks.
  • Ghaemi, M. S., Agard, B., Partovi Nia, V., and Trepanier, M. (2015) Identifying Temporal User Behavior through Smart Card Data, Conference on Advanced Systems in Public Transport (CASPT), to appear.
  • We discuss different methods for analyzing temporal data in public transport.
  • Tilouche, S., Bassetto, S., and Partovi Nia, V. (2014) Classification Algorithms for Virtual Metrology, Proceedings of IEEE International Conference on Management of Innovation and Technology (ICMIT), 455-499
  • We reviewed algorithms that could be applied as a tool in virtual metrology. Our simulation shows the neural networks regression is a strong candidate.
  • Mouret, G., Brault, J.-J., and Partovi Nia, V. (2013) Generalized Elastic Net Regression, JSM Proceedings, Section on Statistical Learning and Data Mining, Montreal, American Statistical Association, 3457-3464.
  • Elastic Net penalty is an effective tool for variable selection in linear regression. This work generalizes the elastic net regularization penalty. We suggest a Bayesian perspective for estimation of the regularization constant.
  • Partovi Nia, V., Asgharian, M., and Bassetto, S. (2013) A Formal Test for Binary R&R Measurement Systems, JSM Proceedings, SSC Section, Montreal, American Statistical Association, 2865-2872
  • Repeatability and Reproducability, called R&R, are minimum requirments of a measurement system. A test of R&R is often required as an improtant part of statistical process control. Here we show that Pearson statistic can be used in order to build a formal statistical test of repeatability and reproducability for a pass-fail inspection measurement.
  • Trepanier, M., Partovi Nia, V. and Agard, B. (2013) Assessing Public Transport Travel Behaviour from Smart Card Data with Advanced Data Mining Techniques, 13th World Conference on Transport Research (WCTR)
  • We discuss various challenges in public transport data analysis. We show various handy tricks to analyze complex public transport data.
  • Drikvandi, R., Verbeke, G., Khodadadi, A, and Partovi Nia, V. (2013) Testing Multiple Variance Components in Linear Mixed Effects Models, Biostatistics, Vol 14, Issue 1, 144-159.
  • Testing multiple variance components is a non-standard, important, and a difficult problem with a lot of applications in biology, medicine, ecology and many others. Here we propose a permutation test based on a simple test statistic. This work introduces a methodology for testing even a subset of variance components with zero.
  • Partovi Nia, V. and Davison A. (2012) High-Dimensional Bayesian Clustering with Variable Selection: The R Package bclust, Jounral of Statistical Software, Vol 47, Issue 5, 1-22.
  • We show how Bayesian clustering with variable selection can be done using our package bclust. The bclust package implements a new Bayesian framework for variable selection in high-dimensional clustering.
  • Partovi Nia, V. and Stephens D. A. (2011) Dendrogram Representation of Stochastic Clustering, Probability: Interpretation, Theory and Applications, Nova Publishers, Chapter 6, 203-217. Buy the book [HERE] or from [AMAZON].
  • We show how a hierarchical tree can be extracted from random samples of groupings. We also explain the ideas using one of our R packages published on CRAN.
  • Partovi Nia, V. (2009) Fast High-Dimensional Bayesian Classification and Clustering, Ecole Polytechnique Federale de Lausanne, Ph.D. Thesis.
  • Parchami, A., Mashinchi, M., and Partovi Nia, V. (2008) A Consistent Confidence Interval for Fuzzy Capability Index, Applied and Computational Mathematics, Vol. 7, no. 1, 119-125.
  • Messerli, G., Partovi Nia, V., Trevisan, M., Kolbe, A., Schauer, N., Geigenberger, P., Chen, J., Davison, A. C., Fernie, A. R. and Zeeman, S. C. (2007) Rapid Classification of Phenotypic Mutants of Arabidopsis Via Metabolite Fingerprinting, Plant Physiology, Vol. 143, 1484-1492.
  • We show metabolite fingerprinting is an effective method of classification forward genetic mutants of Arbidopsis Thaliana.
  • Partovi Nia, V. (2006) Gauss-Hermite Quadratures: Numerical or Statistical Method? The 8th. Iranian Statistical Conference Proceedings (invited and refereed papers), 209-215