Person: KANDEMİR, Malih
Loading...
Email Address
Birth Date
Name
Job Title
First Name
Malih
Last Name
KANDEMİR
7 results
Publication Search Results
Now showing 1 - 7 of 7
Conference paperPublication Metadata only Sampling-free variational inference of Bayesian neural networks by variance backpropagation(Association For Uncertainty in Artificial Intelligence (AUAI), 2019) Haußmann, M.; Hamprecht, F. A.; Kandemir, Melih; Computer Science; KANDEMİR, MalihWe propose a new Bayesian Neural Net formulation that affords variational inference for which the evidence lower bound is analytically tractable subject to a tight approximation. We achieve this tractability by (i) decomposing ReLU nonlinearities into the product of an identity and a Heaviside step function, (ii) introducing a separate path that decomposes the neural net expectation from its variance. We demonstrate formally that introducing separate latent binary variables to the activations allows representing the neural network likelihood as a chain of linear operations. Performing variational inference on this construction enables a sampling-free computation of the evidence lower bound which is a more effective approximation than the widely applied Monte Carlo sampling and CLT related techniques. We evaluate the model on a range of regression and classification tasks against BNN inference alternatives, showing competitive or improved performance over the current state-of-the-art.Conference paperPublication Metadata only Prediction of active UE number with Bayesian neural networks for self-organizing LTE networks(IEEE, 2017-01-01) Narmanlıoğlu, Ö.; Zeydan, E.; Kandemir, Melih; Kranda, T.; Computer Science; KANDEMİR, MalihInternet-empowered electronic gadgets and content rich multimedia applications have expanded exponentially in recent years. As a consequence, heterogeneous network structures introduced with Long Term Evolution (LTE) Advanced have increasingly gaining momentum in order to handle with data explosion. On the other hand, the deployment of new network equipment is resulting in increasing both capital and operating expenditures. These deployments are done under the consideration of the busy hour periods which the network experiences the highest amount of traffic. However, these periods refer to only a couple of hours over a 24-hour period. In relation to this, accurate prediction of active user equipment (UE) number is significant for efficient network operations and results in decreasing energy consumption. In this paper, we investigate a Bayesian technique to design an optimal feed-forward neural network for shortterm predictor executed at the network management entity and providing proactivity to Energy Saving, a Self-Organizing Network function. We first demonstrate prediction results of active UE number collected from real LTE network. Then, we evaluate the prediction accuracy of the Bayesian neural network as comparing with low complex naive prediction method, Holt- Winter's exponential smoothing method, a deterministic feedforward neural network without Bayesian regularization term.Conference paperPublication Metadata only Sampling-free variational inference of bayesian neural networks by variance backpropagation(ML Research Press, 2020) Haußmann, M.; Hamprecht, F. A.; Kandemir, Melih; Computer Science; KANDEMİR, MalihWe propose a new Bayesian Neural Net formulation that affords variational inference for which the evidence lower bound is analytically tractable subject to a tight approximation. We achieve this tractability by (i) decomposing ReLU nonlinearities into the product of an identity and a Heaviside step function, (ii) introducing a separate path that decomposes the neural net expectation from its variance. We demonstrate formally that introducing separate latent binary variables to the activations allows representing the neural network likelihood as a chain of linear operations. Performing variational inference on this construction enables a sampling-free computation of the evidence lower bound which is a more effective approximation than the widely applied Monte Carlo sampling and CLT related techniques. We evaluate the model on a range of regression and classification tasks against BNN inference alternatives, showing competitive or improved performance over the current state-of-the-art.ArticlePublication Metadata only Variational closed-Form deep neural net inference(Elsevier, 2018-09) Kandemir, Melih; Computer Science; KANDEMİR, MalihWe introduce a Bayesian construction for deep neural networks that is amenable to mean field variational inference that operates solely by closed-form update rules. Hence, it does not require any learning rate to be manually tuned. We show that by this virtue it becomes possible with our model to perform effective deep learning on three setups where conventional neural nets are known to perform suboptimally: i) online learning, ii) learning from small data, and iii) active learning. We compare our approach to earlier Bayesian neural network inference techniques spanning from expectation propagation to gradient-based variational Bayes, as well as deterministic neural nets with various activations functions. We observe our approach to improve on all these alternatives in two mainstream vision benchmarks and two medical data sets: diabetic retinopathy screening and exudate detection from eye fundus images.ArticlePublication Metadata only Supervising topic models with Gaussian processes(Elsevier, 2018-05) Kandemir, Melih; Kekeç, T.; Yeniterzi, Reyyan; Computer Science; KANDEMİR, Malih; YENİTERZİ, ReyyanTopic modeling is a powerful approach for modeling data represented as high-dimensional histograms. While the high dimensionality of such input data is extremely beneficial in unsupervised applications including language modeling and text data exploration, it introduces difficulties in cases where class information is available to boost up prediction performance. Feeding such input directly to a classifier suffers from the curse of dimensionality. Performing dimensionality reduction and classification disjointly, on the other hand, cannot enjoy optimal performance due to information loss in the gap between these two steps unaware of each other. Existing supervised topic models introduced as a remedy to such scenarios have thus far incorporated only linear classifiers in order to keep inference tractable, causing a dramatical sacrifice from expressive power. In this paper, we propose the first Bayesian construction to perform topic modeling and non-linear classification jointly. We use the well-known Latent Dirichlet Allocation (LDA) for topic modeling and sparse Gaussian processes for non-linear classification. We combine these two components by a latent variable encoding the empirical topic distribution of each document in the corpus. We achieve a novel variational inference scheme by adapting ideas from the newly emerging deep Gaussian processes into the realm of topic modeling. We demonstrate that our model outperforms other existing approaches such as: (i) disjoint LDA and non-linear classification, (ii) joint LDA and linear classification, (iii) joint non-LDA linear subspace modeling and linear classification, and (iv) non-linear classification without topic modeling, in three benchmark data sets from two real-world applications: text categorization and image tagging.Conference paperPublication Metadata only On context-aware DDoS attacks using deep generative networks(IEEE, 2018-10) Gürsun, Gonca; Şensoy, Murat; Kandemir, Melih; Computer Science; GÜRSUN, Gonca; ŞENSOY, Murat; KANDEMİR, MalihDistributed Denial of Service (DDoS) attacks continue to be one of the most severe threats in the Internet. The intrinsic challenge in preventing DDoS attacks is to distinguish them from legitimate flash crowds since two have many traffic characteristics in common. Today most DDoS detection techniques focus on finding parametric differences between the patterns in attack and legitimate traffic. However, such techniques are very sensitive to the threshold values set on the parameters and more importantly legitimate traffic features might be mimicked by smart attackers to generate requests that look like flash crowds. In this paper, we propose a framework for training networks for such smart attacks. Our framework is based on Deep Generative Network models and our contributions are two-fold.We first show that legitimate traffic features can be mimicked without explicitly modeling their distributions. Second, we introduce the concept of context-aware DDoS attacks. We show that an attacker can generate traffic that looks similar to flash crowds to be undetected for long periods of time. However, the ability of generating such attacks is constrained by the budget of the attacker. A context-aware attacker is the one that can intelligently use its budget to maximize the damage in the victim network. Our study provides a framework for training networks for such DDoS attack scenarios.Conference paperPublication Metadata only Variational bayesian multiple instance learning with gaussian processes(IEEE, 2017) Haussmann, M.; Hamprecht, F. A.; Kandemir, Melih; Computer Science; KANDEMİR, MalihGaussian Processes (GPs) are effective Bayesian predictors. We here show for the first time that instance labels of a GP classifier can be inferred in the multiple instance learning (MIL) setting using variational Bayes. We achieve this via a new construction of the bag likelihood that assumes a large value if the instance predictions obey the MIL constraints and a small value otherwise. This construction lets us derive the update rules for the variational parameters analytically, assuring both scalable learning and fast convergence. We observe this model to improve the state of the art in instance label prediction from bag-level supervision in the 20 Newsgroups benchmark, as well as in Barretts cancer tumor localization from histopathology tissue microarray images. Furthermore, we introduce a novel pipeline for weakly supervised object detection naturally complemented with our model, which improves the state of the art on the PASCAL VOC 2007 and 2012 data sets. Last but not least, the performance of our model can be further boosted up using mixed supervision: a combination of weak (bag) and strong (instance) labels.