arXiv subjects

Arun Kumar

Publications and source records attributed to Arun Kumar.

At least 91 records · Page 5Linked to original sources

arXiv2018

Deep Domain Adaptation under Deep Label Scarcity

The goal behind Domain Adaptation (DA) is to leverage the labeled examples from a source domain so as to infer an accurate model in a target domain where labels are not available or in scarce at the best. A state-of-the-art approach for the DA is due to (Ganin et al. 2016), known as DANN, where they attempt to induce a common representation of source and target domains via adversarial training. This approach requires a large number of labeled examples from the source domain to be able to infer a good model for the target domain. However, in many situations obtaining labels in the source domain is expensive which results in deteriorated performance of DANN and limits its applicability in such scenarios. In this paper, we propose a novel approach to overcome this limitation. In our work, we first establish that DANN reduces the original DA problem into a semi-supervised learning problem over the space of common representation. Next, we propose a learning approach, namely TransDANN, that amalgamates adversarial learning and transductive learning to mitigate the detrimental impact of limited source labels and yields improved performance. Experimental results (both on text and images) show a significant boost in the performance of TransDANN over DANN under such scenarios. We also provide theoretical justification for the performance boost.

cs.LG↗

arXiv2018

In-RDBMS Hardware Acceleration of Advanced Analytics

The data revolution is fueled by advances in machine learning, databases, and hardware design. Programmable accelerators are making their way into each of these areas independently. As such, there is a void of solutions that enables hardware acceleration at the intersection of these disjoint fields. This paper sets out to be the initial step towards a unifying solution for in-Database Acceleration of Advanced Analytics (DAnA). Deploying specialized hardware, such as FPGAs, for in-database analytics currently requires hand-designing the hardware and manually routing the data. Instead, DAnA automatically maps a high-level specification of advanced analytics queries to an FPGA accelerator. The accelerator implementation is generated for a User Defined Function (UDF), expressed as a part of an SQL query using a Python-embedded Domain-Specific Language (DSL). To realize an efficient in-database integration, DAnA accelerators contain a novel hardware structure, Striders, that directly interface with the buffer pool of the database. Striders extract, cleanse, and process the training data tuples that are consumed by a multi-threaded FPGA engine that executes the analytics algorithm. We integrate DAnA with PostgreSQL to generate hardware accelerators for a range of real-world and synthetic datasets running diverse ML algorithms. Results show that DAnA-enhanced PostgreSQL provides, on average, 8.3x end-to-end speedup for real datasets, with a maximum of 28.2x. Moreover, DAnA-enhanced PostgreSQL is, on average, 4.0x faster than the multi-threaded Apache MADLib running on Greenplum. DAnA provides these benefits while hiding the complexity of hardware design from data scientists and allowing them to express the algorithm in =30-60 lines of Python.

cs.DB↗

arXiv2018

$D$-dimensional Bardeen-AdS black holes in Einstein-Gauss-Bonnet theory

We present a $D$-dimensional Bardeen like Anti-de Sitter (AdS) black hole solution in Einstein-Gauss-Bonnet (EGB) gravity, \textit{viz}., Bardeen-EGB-AdS black holes. The Bardeen-EGB-AdS black hole has an additional parameter due to charge ($e$), apart from mass ($M$) and Gauss-Bonnet parameter ($α$). Interestingly, for each value of $α$, there exist a critical $e = e_E$ which corresponds to an extremal regular black hole with degenerate horizons, while for $e< e_E$, it describes non-extremal black hole with two horizons. Despite the complicated solution, the thermodynamical quantities, like temperature ($T$), specific heat($C$) and entropy ($S$) associated with the black hole are obtained exactly. It turns out that the heat capacity diverges at critical horizon radius $r_+ = r_C$, where the temperature attains maximum value and the Hawking-Page transition is achievable. Thus, we have an exact $D$-dimensional regular black holes, when evaporates lead to a thermodynamical stable remnant.

gr-qc↗

arXiv2018

Tempered Fractional Poisson Processes and Fractional Equations with Z-Transform

In this article, we derive the state probabilities of different type of space- and time-fractional Poisson processes using z-transform. We work on tempered versions of time-fractional Poisson process and space-fractional Poisson processes. We also introduce Gegenbauer type fractional differential equations and their solutions using z-transform. Our results generalize and complement the result available on fractional Poisson processes in several directions.

math.PR↗

arXiv2018

Bipedal Walking Robot using Deep Deterministic Policy Gradient

Machine learning algorithms have found several applications in the field of robotics and control systems. The control systems community has started to show interest towards several machine learning algorithms from the sub-domains such as supervised learning, imitation learning and reinforcement learning to achieve autonomous control and intelligent decision making. Amongst many complex control problems, stable bipedal walking has been the most challenging problem. In this paper, we present an architecture to design and simulate a planar bipedal walking robot(BWR) using a realistic robotics simulator, Gazebo. The robot demonstrates successful walking behaviour by learning through several of its trial and errors, without any prior knowledge of itself or the world dynamics. The autonomous walking of the BWR is achieved using reinforcement learning algorithm called Deep Deterministic Policy Gradient(DDPG). DDPG is one of the algorithms for learning controls in continuous action spaces. After training the model in simulation, it was observed that, with a proper shaped reward function, the robot achieved faster walking or even rendered a running gait with an average speed of 0.83 m/s. The gait pattern of the bipedal walker was compared with the actual human walking pattern. The results show that the bipedal walking pattern had similar characteristics to that of a human walking pattern. The video presenting our experiment is available at https://goo.gl/NHXKqR.

cs.RO↗

arXiv2018

Model-based Pricing for Machine Learning in a Data Marketplace

Data analytics using machine learning (ML) has become ubiquitous in science, business intelligence, journalism and many other domains. While a lot of work focuses on reducing the training cost, inference runtime and storage cost of ML models, little work studies how to reduce the cost of data acquisition, which potentially leads to a loss of sellers' revenue and buyers' affordability and efficiency. In this paper, we propose a model-based pricing (MBP) framework, which instead of pricing the data, directly prices ML model instances. We first formally describe the desired properties of the MBP framework, with a focus on avoiding arbitrage. Next, we show a concrete realization of the MBP framework via a noise injection approach, which provably satisfies the desired formal properties. Based on the proposed framework, we then provide algorithmic solutions on how the seller can assign prices to models under different market scenarios (such as to maximize revenue). Finally, we conduct extensive experiments, which validate that the MBP framework can provide high revenue to the seller, high affordability to the buyer, and also operate on low runtime cost.

cs.DB↗

arXiv2018

Evidence for superferrimagnetic clusters and spin-glass transition involving 4f Dy3+ spins in h-DyMnO3: A new twist to 4f Re3+ spin ordering in hexagonal manganites

The ferroelectric phase of the multiferroic hexagonal manganites (h-ReMnO3) has been reported to undergo a series of magnetic transitions involving long-range ordering/reorientation of 4fRe3+ and/or 3dMn3+ spins below room temperature. These transitions have attracted a lot of attention in recent years due to the geometrically frustrated nature of magnetic interactions. We have revisited these transitions in high quality single crystals of h-DyMnO3 using dc and ac susceptibility measurements as a function of temperature (T), magnetic field (H) and frequency (w) supplemented by specific heat measurements. Taking h-DyMnO3 as an example, we show that the Dy3+ spins below TN~68K are in a superferrimagnetic (SFIM) state whereas they undergo spin-glass (SG) transition below TDy3+~7K. Our observations demonstrate that neither the Néel transition at TN~68K nor the transition at TDy3+~7K is associated with long-range ordered states of Dy3+ spins as believed so far in the literature. The SG state of h-DyMnO3 is quite exotic as it occurs in an ordered compound purely due to geometrical frustration without any random disorder. Further, it shows an interesting crossover from de Almeida-Thouless type exponent (m=2/3) to Gabay-Toulouse type (m=2) with increasing field which cannot be explained in terms of the existing mean field theories of SG transition in Ising or Heisenberg systems but is expected for a vector X-Y SG system. Our observations call for a systematic reinvestigation of the nature of magnetic transitions involving Re3+ ions in other h-ReMnO3 also.

cond-mat.mtrl-sci↗

arXiv2018

Evidence for two spin-glass transitions with magnetoelastic and magnetoelectric couplings in the multiferroic (Bi$_{1-x}$Ba$_x$)(Fe$_{1-x}$Ti$_x$)O$_3$ system

For disordered Heisenberg systems with small single ion anisotropy, two spin glass transitions below the long range ordered phase transition temperature has been predicted theoretically for compositions close to the percolation threshold. Experimental verification of these predictions is still controversial for conventional spin glasses. We show that multiferroic spin glass systems can provide a unique platform for verifying these theoretical predictions via a study of change in magnetoelastic and magnetoelectric couplings, obtained from an analysis of diffraction data, at the spin glass transition temperatures. Results of macroscopic and microscopic (x-ray and neutron scattering) measurements are presented on disordered BiFeO3, a canonical Heisenberg system with small single ion anisotropy, which reveal appearance of two spin glass phases SG1 and SG2 in coexistence with the LRO phase below the A-T and G-T lines. It is shown that the temperature dependence of the integrated intensity of the antiferromagnetic peak shows dips with respect to the Brillouin function behaviour around the SG1 and SG2 transition temperatures. The ferroelectric polarisation changes significantly at the two spin glass transition temperatures. These results, obtained using microscopic techniques, clearly demonstrate that the SG1 and SG2 transitions occur on the same magnetic sublattice and are intrinsic to the system. We also construct a phase diagram showing all the magnetic phases in BF-xBT system. While our results on the two spin glass transitions support the theoretical predictions, it also raises several open questions which need to be addressed by revisiting the existing theories of spin glass transitions by taking into account the effect of magnetoelastic and magnetoelectric couplings as well as electromagnons.

cond-mat.mtrl-sci↗

arXiv2017

Higgs boson measurements in $WW$, $ττ$ and $μμ$ channels with CMS

This note presents search for the Standard Model Higgs boson in $WW$, $ττ$ and $μμ$ decay channels with the proton-proton collision data collected by the CMS experiment at the LHC. The results have been derived using different amounts of luminosities for different channels.

hep-ex↗

arXiv2017

Stone and double Stone algebras: Boolean and Rough Set Representations, 3-valued and 4-valued Logics

Moisil in 1941, while constructing the algebraic models of n-valued Łukasiewicz logic defined the set $B^{[n]}$,where $B$ is a Boolean algebra and `n' being a natural number. Further it was proved by Moisil himself the representations of n-valued Łukasiewicz Moisil algebra in terms of $B^{[n]}$. In this article, structural representation results for Stone, dual Stone and double Stone algebras are proved similar to Moisil's work by showing that elements of these algebras can be looked upon as monotone ordered tuple of sets. 3-valued semantics of logic for Stone algebra, dual Stone algebras and 4-valued semantics of logic for double Stone algebras are proposed and established soundness and completeness results.

math.LO↗

arXiv2017

Dialogue Act Sequence Labeling using Hierarchical encoder with CRF

Dialogue Act recognition associate dialogue acts (i.e., semantic labels) to utterances in a conversation. The problem of associating semantic labels to utterances can be treated as a sequence labeling problem. In this work, we build a hierarchical recurrent neural network using bidirectional LSTM as a base unit and the conditional random field (CRF) as the top layer to classify each utterance into its corresponding dialogue act. The hierarchical network learns representations at multiple levels, i.e., word level, utterance level, and conversation level. The conversation level representations are input to the CRF layer, which takes into account not only all previous utterances but also their dialogue acts, thus modeling the dependency among both, labels and utterances, an important consideration of natural dialogue. We validate our approach on two different benchmark data sets, Switchboard and Meeting Recorder Dialogue Act, and show performance improvement over the state-of-the-art methods by $2.2\%$ and $4.1\%$ absolute points, respectively. It is worth noting that the inter-annotator agreement on Switchboard data set is $84\%$, and our method is able to achieve the accuracy of about $79\%$ despite being trained on the noisy data.

cs.CL↗

arXiv2017

Towards Linear Algebra over Normalized Data

Providing machine learning (ML) over relational data is a mainstream requirement for data analytics systems. While almost all the ML tools require the input data to be presented as a single table, many datasets are multi-table, which forces data scientists to join those tables first, leading to data redundancy and runtime waste. Recent works on "factorized" ML mitigate this issue for a few specific ML algorithms by pushing ML through joins. But their approaches require a manual rewrite of ML implementations. Such piecemeal methods create a massive development overhead when extending such ideas to other ML algorithms. In this paper, we show that it is possible to mitigate this overhead by leveraging a popular formal algebra to represent the computations of many ML algorithms: linear algebra. We introduce a new logical data type to represent normalized data and devise a framework of algebraic rewrite rules to convert a large set of linear algebra operations over denormalized data into operations over normalized data. We show how this enables us to automatically "factorize" several popular ML algorithms, thus unifying and generalizing several prior works. We prototype our framework in the popular ML environment R and an industrial R-over-RDBMS tool. Experiments with both synthetic and real normalized data show that our framework also yields significant speed-ups, up to 36x on real data.

cs.DB↗

arXiv2017

Are Key-Foreign Key Joins Safe to Avoid when Learning High-Capacity Classifiers?

Machine learning (ML) over relational data is a booming area of the database industry and academia. While several projects aim to build scalable and fast ML systems, little work has addressed the pains of sourcing data and features for ML tasks. Real-world relational databases typically have many tables (often, dozens) and data scientists often struggle to even obtain and join all possible tables that provide features for ML. In this context, Kumar et al. showed recently that key-foreign key dependencies (KFKDs) between tables often lets us avoid such joins without significantly affecting prediction accuracy--an idea they called avoiding joins safely. While initially controversial, this idea has since been used by multiple companies to reduce the burden of data sourcing for ML. But their work applied only to linear classifiers. In this work, we verify if their results hold for three popular complex classifiers: decision trees, SVMs, and ANNs. We conduct an extensive experimental study using both real-world datasets and simulations to analyze the effects of avoiding KFK joins on such models. Our results show that these high-capacity classifiers are surprisingly and counter-intuitively more robust to avoiding KFK joins compared to linear classifiers, refuting an intuition from the prior work's analysis. We explain this behavior intuitively and identify open questions at the intersection of data management and ML theoretical research. All of our code and datasets are available for download from http://cseweb.ucsd.edu/~arunkk/hamlet.

cs.DB↗

arXiv2017

Bolt-on Differential Privacy for Scalable Stochastic Gradient Descent-based Analytics

While significant progress has been made separately on analytics systems for scalable stochastic gradient descent (SGD) and private SGD, none of the major scalable analytics frameworks have incorporated differentially private SGD. There are two inter-related issues for this disconnect between research and practice: (1) low model accuracy due to added noise to guarantee privacy, and (2) high development and runtime overhead of the private algorithms. This paper takes a first step to remedy this disconnect and proposes a private SGD algorithm to address \emph{both} issues in an integrated manner. In contrast to the white-box approach adopted by previous work, we revisit and use the classical technique of {\em output perturbation} to devise a novel "bolt-on" approach to private SGD. While our approach trivially addresses (2), it makes (1) even more challenging. We address this challenge by providing a novel analysis of the $L_2$-sensitivity of SGD, which allows, under the same privacy guarantees, better convergence of SGD when only a constant number of passes can be made over the data. We integrate our algorithm, as well as other state-of-the-art differentially private SGD, into Bismarck, a popular scalable SGD-based analytics system on top of an RDBMS. Extensive experiments show that our algorithm can be easily integrated, incurs virtually no overhead, scales well, and most importantly, yields substantially better (up to 4X) test accuracy than the state-of-the-art algorithms on many real datasets.

cs.LG↗

arXiv2017

A note on dual demodulator continuous transmission frequency modulation technique

The range resolution in conventional continuous time frequency modulation (CTFM) is inversely proportional to the signal bandwidth. The dual-demodulator continuous time frequency modulation (DD-CTFM) processing technique was proposed by Gough et al [1] as a method to increase the range resolution by making the output of DD-CTFM truly continuous. However, it has been found that in practice the range resolution is still limited by the signal bandwidth. The limitation of DD-CTFM has been explained using simulations and mathematically in this paper.

cs.IT↗

arXiv2016

Novelty Learning via Collaborative Proximity Filtering

The vast majority of recommender systems model preferences as static or slowly changing due to observable user experience. However, spontaneous changes in user preferences are ubiquitous in many domains like media consumption and key factors that drive changes in preferences are not directly observable. These latent sources of preference change pose new challenges. When systems do not track and adapt to users' tastes, users lose confidence and trust, increasing the risk of user churn. We meet these challenges by developing a model of novelty preferences that learns and tracks latent user tastes. We combine three innovations: a new measure of item similarity based on patterns of consumption co-occurrence; model for {\em spontaneous} changes in preferences; and a learning agent that tracks each user's dynamic preferences and learns individualized policies for variety. The resulting framework adaptively provides users with novelty tailored to their preferences for change per se.

cs.HC↗

arXiv2016

The Ground state of BiFeO3: Low temperature magnetic phase transitions revisited

Recent neutron diffraction and NMR studies suggest that the incommensurately modulated spin cycloid structure of BiFeO3 is stable down to 4.2 K, whereas DC [M(T)] and AC [\c{hi} (ω, T)] magnetization, and caloric studies have revealed several magnetic transitions including a spin glass transition around 25 K. The two sets of observations are irreconcilable and to settle this, it is important to first verify if the low temperature magnetic transitions are intrinsic to BiFeO3 or some of them are offshoots of oxygen vacancies and the associated redox reaction involving conversion of Fe3+ to Fe2+. We present here the results of M (T) and \c{hi} (ω, T) measurements on pure and 0.3 wt% MnO2 doped BiFeO3 samples in the 2 to 300 K temperature range. It is shown that MnO2 doping increases the resistivity of the samples by three orders of magnitude as a result of reduced oxygen vacancy concentration. A comparative study of the M (T) and AC \c{hi} (ω, T) results on two types of samples reveals that the transitions around 25 K, 110 K and 250 K may be intrinsic to BiFeO3. The widely reported transition at 50 K is argued to be defect induced, as it is absent in the doped samples. We also show that the spin glass transition temperature TSG is less than the spin glass freezing temperature (Tf), as expected for both canonical and cluster glasses, in marked contrast to an earlier report of TSG > Tf which is unphysical. We have also observed a cusp corresponding to the spin glass freezing at Tf in ZFC M (T) data not observed so far by previous workers. We argue that the ground state of BiFeO3 consists of the coexistence of the spin glass phase with the long range ordered AFM phase with a cycloidal spin structure.

cond-mat.mtrl-sci↗

arXiv2016

Mittag-Leffler Lévy Processes

In this article, we introduce Mittag-Leffler Lévy process and provide two alternative representations of this process. First, in terms of Laplace transform of the marginal densities and next as a subordinated stochastic process. Both these representations are useful in analyzing the properties of the process. Since integer order moments for this process are not finite, we obtain fractional order moments. Its density function and corresponding Lévy measure density is also obtained. Further, heavy tailed behavior of densities and stochastic self-similarity of the process is demonstrated. Our results generalize and complement the results available on Mittag-Leffler distribution in several directions.

math.PR↗