no code implementations • 13 Feb 2023 • David Bosch, Ashkan Panahi, Babak Hassibi
We provide exact asymptotic expressions for the performance of regression by an $L-$layer deep random feature (RF) model, where the input is mapped through multiple random embedding and non-linear activation functions.
1 code implementation • 28 Nov 2022 • Fengyu Zhang, Ashkan Panahi, Guangjun Gao
By ablation study, we show that low frequency self-attention can achieve very close or better performance relative to full frequency even without retraining the network.
no code implementations • 2 Oct 2022 • Siddharth Roheda, Ashkan Panahi, Hamid Krim
This is achieved by minimizing the cost of transporting the samples from the target domain to the distribution of the source domain.
no code implementations • 22 Jun 2022 • Lena Stempfle, Ashkan Panahi, Fredrik D. Johansson
Conversely, fitting a single shared model to the full data set relies on imputation which often leads to biased results when missingness depends on unobserved factors.
no code implementations • 6 Apr 2022 • David Bosch, Ashkan Panahi, Ayca Özcelikkale, Devdatt Dubhash
We compute precise asymptotic expressions for the learning curves of least squares random feature (RF) models with either a separable strongly convex regularization or the $\ell_1$ regularization.
no code implementations • 18 Jun 2020 • Sally Ghanem, Ashkan Panahi, Hamid Krim, Ryan A. Kerekes
Robust Subspace Recovery (RoSuRe) algorithm was recently introduced as a principled and numerically efficient algorithm that unfolds underlying Unions of Subspaces (UoS) structure, present in the data.
no code implementations • 30 Mar 2020 • Arman Rahbar, Ashkan Panahi, Chiranjib Bhattacharyya, Devdatt Dubhashi, Morteza Haghir Chehreghani
Knowledge transfer is shown to be a very successful technique for training neural classifiers: together with the ground truth data, it uses the "privileged information" (PI) obtained by a "teacher" network to train a "student" network.
no code implementations • 23 Sep 2019 • Yuming Huang, Ashkan Panahi, Hamid Krim, Liyi Dai
We further demonstrate by numerical experiments that in the presence of consistent communities between different layers, consistent communities are matched, and the detectability is improved over a single layer.
no code implementations • 13 Sep 2019 • Yuming Huang, Ashkan Panahi, Hamid Krim, Yiyi Yu, Spencer L. Smith
We present a novel adversarial framework for training deep belief networks (DBNs), which includes replacing the generator network in the methodology of generative adversarial networks (GANs) with a DBN and developing a highly parallelizable numerical algorithm for training the resulting architecture in a stochastic manner.
no code implementations • 20 Aug 2019 • Shahin Mahdizadehaghdam, Ashkan Panahi, Hamid Krim
To that end we start by dividing an image into multiple patches and modifying the role of the generative network from producing an entire image, at once, to creating a sparse representation vector for each image patch.
no code implementations • 13 Jun 2019 • Wen Tang, Ashkan Panahi, Hamid Krim
Zero-shot learning (ZSL) which aims to recognize unseen object classes by only training on seen object classes, has increasingly been of great interest in Machine Learning, and has registered with some successes.
no code implementations • 9 Mar 2019 • Arman Rahbar, Ashkan Panahi, Morteza Haghir Chehreghani, Devdatt Dubhashi, Hamid Krim
We develop a novel theoretical framework for understating OT schemes respecting a class structure.
no code implementations • 7 Mar 2019 • Wen Tang, Ashkan Panahi, Hamid Krim, Liyi Dai
Discriminative Dictionary Learning (DL) methods have been widely advocated for image classification problems.
1 code implementation • 13 Jul 2018 • Wen Tang, Ashkan Panahi, Hamid Krim, Liyi Dai
A consensus structured analysis dictionary and a global classifier are jointly learned in the distributed approach to safeguard the discriminative power and the efficiency of classification.
1 code implementation • 2 May 2018 • Wen Tang, Ashkan Panahi, Hamid Krim, Liyi Dai
We propose a computationally efficient and high-performance classification algorithm by incorporating class structural information in analysis dictionary learning.
no code implementations • 22 Mar 2018 • Ashkan Panahi, Hamid Krim, Liyi Dai
Parametric approaches to Learning, such as deep learning (DL), are highly popular in nonlinear regression, in spite of their extremely difficult training with their increasing complexity (e. g. number of layers in DL).
no code implementations • 11 Mar 2018 • Shahin Mahdizadehaghdam, Ashkan Panahi, Hamid Krim, Liyi Dai
Deep dictionary learning seeks multiple dictionaries at different image scales to capture complementary coherent characteristics.
no code implementations • NeurIPS 2017 • Ashkan Panahi, Babak Hassibi
Precise expressions for the asymptotic performance of LASSO have been obtained for a number of different cases, in particular when the elements of the dictionary matrix are sampled independently from a Gaussian distribution.
no code implementations • ICML 2017 • Ashkan Panahi, Devdatt Dubhashi, Fredrik D. Johansson, Chiranjib Bhattacharyya
Standard clustering methods such as K-means, Gaussian mixture models, and hierarchical clustering are beset by local minima, which are sometimes drastically suboptimal.