Showing 1–30 of 30 results for author: Roy, B

$\mathbb{A}^1$-homotopy type of $\mathbb{A}^2 \setminus \left\{(0,0) \right\}$

Authors: Utsav Choudhury, Biman Roy

Abstract: In this article we prove that any $\mathbb{A}^1$-connected smooth $k$-variety is $\mathbb{A}^1$-uniruled for any algebraically closed field $k$. We establish that if a non empty open subscheme $X$ of a smooth affine $k$-scheme is $\mathbb{A}^1$-weakly equivalent to $\mathbb{A}^2_{k} \setminus \left\{(0,0) \right\}$, then $X \cong \mathbb{A}^2_{k} \setminus \left\{(0,0) \right\}$ as $k$-varieties f… ▽ More In this article we prove that any $\mathbb{A}^1$-connected smooth $k$-variety is $\mathbb{A}^1$-uniruled for any algebraically closed field $k$. We establish that if a non empty open subscheme $X$ of a smooth affine $k$-scheme is $\mathbb{A}^1$-weakly equivalent to $\mathbb{A}^2_{k} \setminus \left\{(0,0) \right\}$, then $X \cong \mathbb{A}^2_{k} \setminus \left\{(0,0) \right\}$ as $k$-varieties for any field $k$ of characteristic $0$. △ Less

Submitted 1 April, 2024; originally announced April 2024.

MSC Class: 14F42; 19E15

Elliptic curve over totally real fields: A Survey

Authors: Bidisha Roy, Lalit Vaishya

Abstract: In this survey article, we summarise the known results towards the conjecture: elliptic curves over totally real number fields are modular. For understanding these recent results in the literature, we present some necessary background along with certain applications. In this survey article, we summarise the known results towards the conjecture: elliptic curves over totally real number fields are modular. For understanding these recent results in the literature, we present some necessary background along with certain applications. △ Less

Submitted 18 April, 2023; originally announced April 2023.

Comments: This is a survey for Bhaskaracharya Pratishthana (BP) program proceedings. Any comments are welcome

On Continuous Full-Order Integral-Terminal Sliding Mode Control with Unknown Apriori Bound on Uncertainty

Authors: Jit Koley, Dinesh Patra, Binoy Krishna Roy

Abstract: This study aims at providing a solution to the problem of designing a continuous and finite-time control for a class of nonlinear systems in the presence of matched uncertainty with an unknown apriori bound. First, we propose a Full-Order Integral-Terminal Sliding Manifold (FOITSM) with a conventional (discontinuous) sliding mode to show that it provides the combined attributes of the nonsingular… ▽ More This study aims at providing a solution to the problem of designing a continuous and finite-time control for a class of nonlinear systems in the presence of matched uncertainty with an unknown apriori bound. First, we propose a Full-Order Integral-Terminal Sliding Manifold (FOITSM) with a conventional (discontinuous) sliding mode to show that it provides the combined attributes of the nonsingular terminal and integral sliding mode algorithms. Secondly, an Adaptive Disturbance Observer (ADO) has been designed to alleviate the effect of the uncertainty acting on the system. On application of the ADO-based Full-Order Integral-Terminal Sliding Mode Control (FOITSMC), the chattering phenomenon in control input has been reduced substantially in the presence of conditionally known matched disturbances. Moreover, the adaptive gains of ADO are updated non-monotonically without over-bounding the acting disturbance, yet sustain the global boundedness of state trajectories within a specific bound. %Finally, an application of the proposed algorithm for attitude stabilization of a rigid spacecraft has been successively shown. △ Less

Submitted 20 March, 2024; v1 submitted 5 April, 2023; originally announced April 2023.

Comments: 14 pages, 9 figures

Transcendence measure of $e^{1/n}$

Authors: Marta Dujella, Anne-Maria Ernvall-Hytönen, Linda Frey, Bidisha Roy

Abstract: For a given transcendental number $ξ$ and for any polynomial $P(X)=: λ_0+\cdots+λ_k X^k \in \mathbb{Z}[X]$, we know that $ P(ξ) \neq 0.$ Let $k \geq 1$ and $ω(k, H)$ be the infimum of the numbers $r > 0$ satisfying the estimate $$ \left|λ_0+λ_1 ξ+λ_2 ξ^{2}+ \ldots +λ_kξ^{k}\right| > \frac{1}{H^r}, $$ for all $(λ_0, \ldots ,λ_k)^T \in \mathbb{Z}^{k+1}\setminus\{\overline{0}\}$ with… ▽ More For a given transcendental number $ξ$ and for any polynomial $P(X)=: λ_0+\cdots+λ_k X^k \in \mathbb{Z}[X]$, we know that $ P(ξ) \neq 0.$ Let $k \geq 1$ and $ω(k, H)$ be the infimum of the numbers $r > 0$ satisfying the estimate $$ \left|λ_0+λ_1 ξ+λ_2 ξ^{2}+ \ldots +λ_kξ^{k}\right| > \frac{1}{H^r}, $$ for all $(λ_0, \ldots ,λ_k)^T \in \mathbb{Z}^{k+1}\setminus\{\overline{0}\}$ with $\max_{1\le i\le k} \{|λ_i|\} \le H$. Any function greater than or equal to $ω(k, H)$ is a {\it transcendence measure of $ξ$}. In this article, we find out a transcendence measure of $ e^{1/n}$ which improves a result proved by Mahler(\cite{Mahler}) in 1975. △ Less

Submitted 9 March, 2023; originally announced March 2023.

Comments: This collaboration has come up due to the wonderful programme: Women in Numbers Europe 4

Leveraging Demonstrations to Improve Online Learning: Quality Matters

Authors: Botao Hao, Rahul Jain, Tor Lattimore, Benjamin Van Roy, Zheng Wen

Abstract: We investigate the extent to which offline demonstration data can improve online learning. It is natural to expect some improvement, but the question is how, and by how much? We show that the degree of improvement must depend on the quality of the demonstration data. To generate portable insights, we focus on Thompson sampling (TS) applied to a multi-armed bandit as a prototypical online learning… ▽ More We investigate the extent to which offline demonstration data can improve online learning. It is natural to expect some improvement, but the question is how, and by how much? We show that the degree of improvement must depend on the quality of the demonstration data. To generate portable insights, we focus on Thompson sampling (TS) applied to a multi-armed bandit as a prototypical online learning algorithm and model. The demonstration data is generated by an expert with a given competence level, a notion we introduce. We propose an informed TS algorithm that utilizes the demonstration data in a coherent way through Bayes' rule and derive a prior-dependent Bayesian regret bound. This offers insight into how pretraining can greatly improve online performance and how the degree of improvement increases with the expert's competence level. We also develop a practical, approximate informed TS algorithm through Bayesian bootstrap** and show substantial empirical regret reduction through experiments. △ Less

Submitted 17 May, 2023; v1 submitted 7 February, 2023; originally announced February 2023.

Comments: Accepted at ICML 2023

Frobenius constants for families of elliptic curves

Authors: Bidisha Roy, Masha Vlasenko

Abstract: The paper deals with a class of periods, Frobenius constants, which describe monodromy of Frobenius solutions of differential equations arising in algebraic geometry. We represent Frobenius constants related to families of elliptic curves as iterated integrals of modular forms. Using the theory of periods of modular forms, we then witness some of these constants in terms of zeta values. The paper deals with a class of periods, Frobenius constants, which describe monodromy of Frobenius solutions of differential equations arising in algebraic geometry. We represent Frobenius constants related to families of elliptic curves as iterated integrals of modular forms. Using the theory of periods of modular forms, we then witness some of these constants in terms of zeta values. △ Less

Submitted 22 May, 2023; v1 submitted 30 June, 2022; originally announced June 2022.

Comments: In the updated version, there are some minor changes

MSC Class: 11F11

Maximum Cut on Interval Graphs of Interval Count Two is NP-complete

Authors: Alexey Barsukov, Bodhayan Roy

Abstract: An interval graph has interval count $\ell$ if it has an interval model, where among every $\ell+1$ intervals there are two that have the same length. Maximum Cut on interval graphs has been found to be NP-complete recently by Adhikary et al. while deciding its complexity on unit interval graphs (graphs with interval count one) remains a longstanding open problem. More recently, de Figueiredo et a… ▽ More An interval graph has interval count $\ell$ if it has an interval model, where among every $\ell+1$ intervals there are two that have the same length. Maximum Cut on interval graphs has been found to be NP-complete recently by Adhikary et al. while deciding its complexity on unit interval graphs (graphs with interval count one) remains a longstanding open problem. More recently, de Figueiredo et al. have made an advancement by showing that the problem remains NP-complete on interval graphs of interval count four. In this paper, we show that Maximum Cut is NP-complete even on interval graphs of interval count two. △ Less

Submitted 24 April, 2024; v1 submitted 13 March, 2022; originally announced March 2022.

MSC Class: 35A01; 65L10; 65L12; 65L20; 65L70

$\mathbb{A}^1$-connected components and characterisation of $\mathbb{A}^2$

Authors: Utsav Choudhury, Biman Roy

Abstract: In this article we prove that any $\mathbb{A}^1$-contractible smooth complex surface is isomorphic as a variety to $\mathbb{C}^2$. We show that the $\mathbb{A}^1$-connected component of a variety $X$ contains the information about $\mathbb{A}^1$'s in $X$. In this article we prove that any $\mathbb{A}^1$-contractible smooth complex surface is isomorphic as a variety to $\mathbb{C}^2$. We show that the $\mathbb{A}^1$-connected component of a variety $X$ contains the information about $\mathbb{A}^1$'s in $X$. △ Less

Submitted 8 November, 2023; v1 submitted 24 December, 2021; originally announced December 2021.

Comments: Proof of almost all the main theorems remain same. Corrected several typos and citations. Changed the title. Theorem 1.1 is true now in characteristic 0

MSC Class: 14F42

Moments of Gaussian hypergeometric functions over finite fields

Authors: Ankan Pal, Bidisha Roy, Mohammad Sadek

Abstract: We prove explicit formulas for certain first and second moment sums of families of Gaussian hypergeometric functions $_{n+1}F_n$, $n\ge1$, over finite fields with $q$ elements where $q$ is an odd prime. This enables us to find an estimate for the value $_6F_5(1)$. In addition, we evaluate certain second moments of traces of the family of Clausen elliptic curves in terms of the value $_3F_2(-1)$. T… ▽ More We prove explicit formulas for certain first and second moment sums of families of Gaussian hypergeometric functions $_{n+1}F_n$, $n\ge1$, over finite fields with $q$ elements where $q$ is an odd prime. This enables us to find an estimate for the value $_6F_5(1)$. In addition, we evaluate certain second moments of traces of the family of Clausen elliptic curves in terms of the value $_3F_2(-1)$. These formulas also allow us to express the product of certain $_2F_1$ and $_{n+1}F_n$ functions in terms of finite field Appell series which generalizes current formulas for products of $_2F_1$ functions. We finally give closed form expressions for sums of Gaussian hypergeometric functions defined using different multiplicative characters. △ Less

Submitted 7 February, 2022; v1 submitted 16 November, 2021; originally announced November 2021.

Torsion groups of Mordell curves over number fields of higher degree

Authors: Tomislav Gužvić, Bidisha Roy

Abstract: Mordell curves over a number field $K$ are elliptic curves of the form $ y^2 = x^3 + c$, where $c \in K \setminus \{ 0 \}$. Let $p \geq 5$ be a prime number, $K$ a number field such that $[K:\mathbb{Q}] \in \{ 2p, 3p \}$ and let $E$ be a Mordell curve defined over $K$. We classify all the possible torsion subgroups $E(K)_{\text{tors}}$ for all Mordell curves $E$ defined over $\mathbb{Q}$ when… ▽ More Mordell curves over a number field $K$ are elliptic curves of the form $ y^2 = x^3 + c$, where $c \in K \setminus \{ 0 \}$. Let $p \geq 5$ be a prime number, $K$ a number field such that $[K:\mathbb{Q}] \in \{ 2p, 3p \}$ and let $E$ be a Mordell curve defined over $K$. We classify all the possible torsion subgroups $E(K)_{\text{tors}}$ for all Mordell curves $E$ defined over $\mathbb{Q}$ when $[K: \mathbb{Q}] \in \{2p, 3p \}$. △ Less

Submitted 11 May, 2021; originally announced May 2021.

Comments: 11 pages

Hypermodels for Exploration

Authors: Vikranth Dwaracherla, Xiuyuan Lu, Morteza Ibrahimi, Ian Osband, Zheng Wen, Benjamin Van Roy

Abstract: We study the use of hypermodels to represent epistemic uncertainty and guide exploration. This generalizes and extends the use of ensembles to approximate Thompson sampling. The computational cost of training an ensemble grows with its size, and as such, prior work has typically been limited to ensembles with tens of elements. We show that alternative hypermodels can enjoy dramatic efficiency gain… ▽ More We study the use of hypermodels to represent epistemic uncertainty and guide exploration. This generalizes and extends the use of ensembles to approximate Thompson sampling. The computational cost of training an ensemble grows with its size, and as such, prior work has typically been limited to ensembles with tens of elements. We show that alternative hypermodels can enjoy dramatic efficiency gains, enabling behavior that would otherwise require hundreds or thousands of elements, and even succeed in situations where ensemble methods fail to learn regardless of size. This allows more accurate approximation of Thompson sampling as well as use of more sophisticated exploration schemes. In particular, we consider an approximate form of information-directed sampling and demonstrate performance gains relative to Thompson sampling. As alternatives to ensembles, we consider linear and neural network hypermodels, also known as hypernetworks. We prove that, with neural network base models, a linear hypermodel can represent essentially any distribution over functions, and as such, hypernetworks are no more expressive. △ Less

Submitted 12 June, 2020; originally announced June 2020.

Comments: Published as a conference paper at ICLR 2020

Complexity of Maximum Cut on Interval Graphs

Authors: Ranendu Adhikary, Kaustav Bose, Satwik Mukherjee, Bodhayan Roy

Abstract: We resolve the longstanding open problem concerning the computational complexity of Max Cut on interval graphs by showing that it is NP-complete. We resolve the longstanding open problem concerning the computational complexity of Max Cut on interval graphs by showing that it is NP-complete. △ Less

Submitted 31 March, 2021; v1 submitted 29 May, 2020; originally announced June 2020.

Provably Efficient Reinforcement Learning with Aggregated States

Authors: Shi Dong, Benjamin Van Roy, Zhengyuan Zhou

Abstract: We establish that an optimistic variant of Q-learning applied to a fixed-horizon episodic Markov decision process with an aggregated state representation incurs regret $\tilde{\mathcal{O}}(\sqrt{H^5 M K} + εHK)$, where $H$ is the horizon, $M$ is the number of aggregate states, $K$ is the number of episodes, and $ε$ is the largest difference between any pair of optimal state-action values associate… ▽ More We establish that an optimistic variant of Q-learning applied to a fixed-horizon episodic Markov decision process with an aggregated state representation incurs regret $\tilde{\mathcal{O}}(\sqrt{H^5 M K} + εHK)$, where $H$ is the horizon, $M$ is the number of aggregate states, $K$ is the number of episodes, and $ε$ is the largest difference between any pair of optimal state-action values associated with a common aggregate state. Notably, this regret bound does not depend on the number of states or actions and indicates that asymptotic per-period regret is no greater than $ε$, independent of horizon. To our knowledge, this is the first such result that applies to reinforcement learning with nontrivial value function approximation without any restrictions on transition probabilities. △ Less

Submitted 19 February, 2020; v1 submitted 13 December, 2019; originally announced December 2019.

Torsion groups of Mordell curves over cubic and sextic fields

Authors: Pallab Kanti Dey, Bidisha Roy

Abstract: In this paper, we classify torsion groups of rational Mordell curves explicitly over cubic fields as well as over sextic fields. Also, we classify torsion groups of Mordell curves over cubic fields and for Mordell curves over sextic fields, we produce all possible torsion groups. In this paper, we classify torsion groups of rational Mordell curves explicitly over cubic fields as well as over sextic fields. Also, we classify torsion groups of Mordell curves over cubic fields and for Mordell curves over sextic fields, we produce all possible torsion groups. △ Less

Submitted 21 August, 2019; originally announced August 2019.

On Colourability of Polygon Visibility Graphs

Authors: Onur Çağirici, Petr Hliněný, Bodhayan Roy

Abstract: We study the problem of colouring visibility graphs of polygons. In particular, for visibility graphs of simple polygons, we provide a polynomial algorithm for 4-colouring, and prove that the 5-colourability question is already NP-complete for them. For visibility graphs of polygons with holes, we prove that the 4-colourability question is NP-complete. We study the problem of colouring visibility graphs of polygons. In particular, for visibility graphs of simple polygons, we provide a polynomial algorithm for 4-colouring, and prove that the 5-colourability question is already NP-complete for them. For visibility graphs of polygons with holes, we prove that the 4-colourability question is NP-complete. △ Less

Submitted 5 June, 2019; originally announced June 2019.

MSC Class: 05C15; 68Q25; 68R10; 68U05

Distribution of residues modulo $p$ using the Dirichlet's class number formula

Authors: Jaitra Chattopadhyay, Bidisha Roy, Subha Sarkar, R. Thangadurai

Abstract: Let $p$ be an odd prime number. In this article, we study the number of quadratic residues and non-residues modulo $p$ which are multiples of $2$ or $3$ or $4$ and lying in the interval $[1, p-1]$, by applying the Dirichlet's class number formula for the imaginary quadratic field $\mathbb{Q}(\sqrt{-p})$. Let $p$ be an odd prime number. In this article, we study the number of quadratic residues and non-residues modulo $p$ which are multiples of $2$ or $3$ or $4$ and lying in the interval $[1, p-1]$, by applying the Dirichlet's class number formula for the imaginary quadratic field $\mathbb{Q}(\sqrt{-p})$. △ Less

Submitted 29 January, 2019; v1 submitted 29 September, 2018; originally announced October 2018.

Comments: To appear in the conference proceedings of ICCGNFRT-2017

Quadratic non-residues and non-primitive roots satisfying a coprimality condition

Authors: Jaitra Chattopadhyay, Bidisha Roy, Subha Sarkar, R. Thangadurai

Abstract: Let $q\geq 1$ be any integer and let $ ε\in [\frac{1}{11}, \frac{1}{2})$ be a given real number. In this short note, we prove that for all primes $p$ satisfying $$ p\equiv 1\pmod{q}, \quad \log\log p > \frac{\log 6.83}{\frac{1}{2}-ε} \mbox{ and } \frac{φ(p-1)}{p-1} \leq \frac{1}{2} - ε, $$ there exists a quadratic non-residue $g$ which is not a primitive root modulo $p$ such that… ▽ More Let $q\geq 1$ be any integer and let $ ε\in [\frac{1}{11}, \frac{1}{2})$ be a given real number. In this short note, we prove that for all primes $p$ satisfying $$ p\equiv 1\pmod{q}, \quad \log\log p > \frac{\log 6.83}{\frac{1}{2}-ε} \mbox{ and } \frac{φ(p-1)}{p-1} \leq \frac{1}{2} - ε, $$ there exists a quadratic non-residue $g$ which is not a primitive root modulo $p$ such that $gcd\left(g, \frac{p-1}{q}\right) = 1$. △ Less

Submitted 13 September, 2018; originally announced September 2018.

Comments: to appear in Bulletin of the Australian Mathematical Society

On determination of Zero-sum $\ell$-generalized Schur Numbers for some linear equations

Authors: Bidisha Roy, Subha Sarkar

Abstract: Let $r$, $m$ and $k\geq 2$ be positive integers such that $r\mid k$ and let $v \in \left[ 0,\lfloor \frac{k-1}{2r} \rfloor \right]$ be any integer. For any integer $\ell \in [1, k]$ and $ε\in \{0,1\}$, we let $\mathcal{E}_{v}^{(\ell, ε)}$ be the linear homogeneous equation defined by $\mathcal{E}_{v}^{(\ell, ε)}: x_1 + \cdots + x_{k-(rv+ε)} =x_{k-(rv+ε-1)} +\cdots+ \ell x_{k}$. We denote the numbe… ▽ More Let $r$, $m$ and $k\geq 2$ be positive integers such that $r\mid k$ and let $v \in \left[ 0,\lfloor \frac{k-1}{2r} \rfloor \right]$ be any integer. For any integer $\ell \in [1, k]$ and $ε\in \{0,1\}$, we let $\mathcal{E}_{v}^{(\ell, ε)}$ be the linear homogeneous equation defined by $\mathcal{E}_{v}^{(\ell, ε)}: x_1 + \cdots + x_{k-(rv+ε)} =x_{k-(rv+ε-1)} +\cdots+ \ell x_{k}$. We denote the number $S_{\mathfrak{z},m}^{(\ell, ε)}(k;r;v)$, which is defined to be the least positive integer $t$ such that for any $m$-coloring $χ: [1, t] \to \{0, 1,\ldots,m-1\}$, there exists a solution $(\hat{x}_1, \hat{x}_2, \ldots, \hat{x}_k)$ to the equation $\mathcal{E}_{v}^{(\ell,ε)}$ that satisfies the $r$-zero-sum condition, namely, $\displaystyle\sum_{i=1}^kχ(\hat{x}_i) \equiv 0\pmod{r}$. In this article, we completely determine the constant $S_{\mathfrak{z}, 2}^{(k,1)}(k;r;0)$, $S_{\mathfrak{z}, m}^{(k-1,1)}(k;r;0)$, $S_{\mathfrak{z}, 2}^{(1,1)}(k;2;1)$ and $S_{\mathfrak{z}, r}^{(1,0)}(k;r;v)$. Also, we prove upper bound for the constants $S_{\mathfrak{z},2}^{(2,1)}(k;2;0)$ and $S_{\mathfrak{z},2}^{(1,1)}(k;2;v)$. △ Less

Submitted 27 August, 2018; originally announced August 2018.

On Fractionally Dense Sets

Authors: Jaitra Chattopadhyay, Bidisha Roy, Subha Sarkar

Abstract: In this article, we prove some subsets of the set of natural numbers $\mathbb{N}$ and any non-zero ideals of an order of imaginary quadratic fields are fractionally dense in $\mathbb{R}_{>0}$ and $\mathbb{C}$ respectively. In this article, we prove some subsets of the set of natural numbers $\mathbb{N}$ and any non-zero ideals of an order of imaginary quadratic fields are fractionally dense in $\mathbb{R}_{>0}$ and $\mathbb{C}$ respectively. △ Less

Submitted 30 September, 2018; v1 submitted 22 April, 2018; originally announced April 2018.

Comments: To appear in the Rocky Mountain Journal of Mathematics

The Determination of 2-color zero-sum generalized Schur Numbers

Authors: Aaron Robertson, Bidisha Roy, Subha Sarkar

Abstract: Consider the equation $\mathcal{E}: x_1+ \cdots+x_{k-1} =x_{k}$ and let $k$ and $r$ be positive integers such that $r\mid k$. The number $S_{\mathfrak{z},2}(k;r)$ is defined to be the least positive integer $t$ such that for any 2-coloring $χ: [1, t] \to \{0, 1\}$ there exists a solution $(\hat{x}_1, \hat{x}_2, \ldots, \hat{x}_k)$ to the equation $\mathcal{E}$ satisfying… ▽ More Consider the equation $\mathcal{E}: x_1+ \cdots+x_{k-1} =x_{k}$ and let $k$ and $r$ be positive integers such that $r\mid k$. The number $S_{\mathfrak{z},2}(k;r)$ is defined to be the least positive integer $t$ such that for any 2-coloring $χ: [1, t] \to \{0, 1\}$ there exists a solution $(\hat{x}_1, \hat{x}_2, \ldots, \hat{x}_k)$ to the equation $\mathcal{E}$ satisfying $\displaystyle \sum_{i=1}^kχ(\hat{x}_i) \equiv 0\pmod{r}$. In a recent paper, the first author posed the question of determining the exact value of $S_{\mathfrak{z}, 2}(k;4)$. In this article, we solve this problem and show, more generally, that $S_{\mathfrak{z}, 2}(k, r)=kr - 2r+1$ for all positive integers $k$ and $r$ with $k>r$ and $r \mid k$. △ Less

Submitted 8 March, 2018; v1 submitted 2 March, 2018; originally announced March 2018.

Gaussian-Dirichlet Posterior Dominance in Sequential Learning

Authors: Ian Osband, Benjamin Van Roy

Abstract: We consider the problem of sequential learning from categorical observations bounded in [0,1]. We establish an ordering between the Dirichlet posterior over categorical outcomes and a Gaussian posterior under observations with N(0,1) noise. We establish that, conditioned upon identical data with at least two observations, the posterior mean of the categorical distribution will always second-order… ▽ More We consider the problem of sequential learning from categorical observations bounded in [0,1]. We establish an ordering between the Dirichlet posterior over categorical outcomes and a Gaussian posterior under observations with N(0,1) noise. We establish that, conditioned upon identical data with at least two observations, the posterior mean of the categorical distribution will always second-order stochastically dominate the posterior mean of the Gaussian distribution. These results provide a useful tool for the analysis of sequential learning under categorical outcomes. △ Less

Submitted 9 February, 2018; v1 submitted 14 February, 2017; originally announced February 2017.

Four-connected triangulations of planar point sets

Authors: Ajit Arvind Diwan, Subir Kumar Ghosh, Bodhayan Roy

Abstract: In this paper, we consider the problem of determining in polynomial time whether a given planar point set $P$ of $n$ points admits 4-connected triangulation. We propose a necessary and sufficient condition for recognizing $P$, and present an $O(n^3)$ algorithm of constructing a 4-connected triangulation of $P$. Thus, our algorithm solves a longstanding open problem in computational geometry and ge… ▽ More In this paper, we consider the problem of determining in polynomial time whether a given planar point set $P$ of $n$ points admits 4-connected triangulation. We propose a necessary and sufficient condition for recognizing $P$, and present an $O(n^3)$ algorithm of constructing a 4-connected triangulation of $P$. Thus, our algorithm solves a longstanding open problem in computational geometry and geometric graph theory. We also provide a simple method for constructing a noncomplex triangulation of $P$ which requires $O(n^2)$ steps. This method provides a new insight to the structure of 4-connected triangulation of point sets. △ Less

Submitted 7 October, 2013; originally announced October 2013.

Comments: 35 pages, 35 figures, 5 references

MSC Class: 68R10 ACM Class: G.2.2; F.2.2

Efficient Reinforcement Learning for High Dimensional Linear Quadratic Systems

Authors: Morteza Ibrahimi, Adel Javanmard, Benjamin Van Roy

Abstract: We study the problem of adaptive control of a high dimensional linear quadratic (LQ) system. Previous work established the asymptotic convergence to an optimal controller for various adaptive control schemes. More recently, for the average cost LQ problem, a regret bound of ${O}(\sqrt{T})$ was shown, apart form logarithmic factors. However, this bound scales exponentially with $p$, the dimension o… ▽ More We study the problem of adaptive control of a high dimensional linear quadratic (LQ) system. Previous work established the asymptotic convergence to an optimal controller for various adaptive control schemes. More recently, for the average cost LQ problem, a regret bound of ${O}(\sqrt{T})$ was shown, apart form logarithmic factors. However, this bound scales exponentially with $p$, the dimension of the state space. In this work we consider the case where the matrices describing the dynamic of the LQ system are sparse and their dimensions are large. We present an adaptive control scheme that achieves a regret bound of ${O}(p \sqrt{T})$, apart from logarithmic factors. In particular, our algorithm has an average cost of $(1+\eps)$ times the optimum cost after $T = \polylog(p) O(1/\eps^2)$. This is in comparison to previous work on the dense dynamics where the algorithm requires time that scales exponentially with dimension in order to achieve regret of $\eps$ times the optimal cost. We believe that our result has prominent applications in the emerging area of computational advertising, in particular targeted online advertising and advertising in social networks. △ Less

Submitted 24 March, 2013; originally announced March 2013.

Comments: 16 pages

Journal ref: Advances in Neural Information Processing Systems (NIPS) 2012: 2645-2653

Infinite families of (non)-Hermitian Hamiltonians associated with exceptional $X_m$ Jacobi polynomials

Authors: Bikashkali Midya, Barnana Roy

Abstract: Using an appropriate change of variable, the Schrödinger equation is transformed into a second-order differential equation satisfied by recently discovered Jacobi type $X_m$ exceptional orthogonal polynomials. This facilitates the derivation of infinite families of exactly solvable Hermitian as well as non-Hermitian trigonometric Scarf potentials and finite number of Hermitian and infinite number… ▽ More Using an appropriate change of variable, the Schrödinger equation is transformed into a second-order differential equation satisfied by recently discovered Jacobi type $X_m$ exceptional orthogonal polynomials. This facilitates the derivation of infinite families of exactly solvable Hermitian as well as non-Hermitian trigonometric Scarf potentials and finite number of Hermitian and infinite number of non-Hermitian PT-symmetric hyperbolic Scarf potentials. The bound state solutions of all these potentials are associated with the aforesaid exceptional orthogonal polynomials. These infinite families of potentials are shown to be extensions of the conventional trigonometric and hyperbolic Scarf potentials by the addition of some rational terms characterized by the presence of classical Jacobi polynomials. All the members of a particular family of these 'rationally extended polynomial-dependent' potentials have the same energy spectrum and possess translational shape invariant symmetry. The obtained non-Hermitian trigonometric Scarf potentials are shown to be quasi-Hermitian in nature ensuring the reality of the associated energy spectra. △ Less

Submitted 12 March, 2013; v1 submitted 29 September, 2012; originally announced October 2012.

Comments: Revised version. Accepted in J. Phys. A

Journal ref: Journal of Physics A 46 (2013) 175201

Some Results On Point Visibility Graphs

Authors: Subir Kumar Ghosh, Bodhayan Roy

Abstract: In this paper, we present three necessary conditions for recognizing point visibility graphs. We show that this recognition problem lies in PSPACE. We state new properties of point visibility graphs along with some known properties that are important in understanding point visibility graphs. For planar point visibility graphs, we present a complete characterization which leads to a linear time rec… ▽ More In this paper, we present three necessary conditions for recognizing point visibility graphs. We show that this recognition problem lies in PSPACE. We state new properties of point visibility graphs along with some known properties that are important in understanding point visibility graphs. For planar point visibility graphs, we present a complete characterization which leads to a linear time recognition and reconstruction algorithm. △ Less

Submitted 4 April, 2014; v1 submitted 11 September, 2012; originally announced September 2012.

MSC Class: 68R10 ACM Class: G.2.2; F.2.2

Combinatorial Batch Codes: A Lower Bound and Optimal Constructions

Authors: Srimanta Bhattacharya, Sushmita Ruj, Bimal Roy

Abstract: Batch codes, introduced by Ishai, Kushilevitz, Ostrovsky and Sahai in [1], are methods for solving the following data storage problem: n data items are to be stored in m servers in such a way that any k of the n items can be retrieved by reading at most t items from each server, and that the total number of items stored in m servers is N . A Combinatorial batch code (CBC) is a batch code where eac… ▽ More Batch codes, introduced by Ishai, Kushilevitz, Ostrovsky and Sahai in [1], are methods for solving the following data storage problem: n data items are to be stored in m servers in such a way that any k of the n items can be retrieved by reading at most t items from each server, and that the total number of items stored in m servers is N . A Combinatorial batch code (CBC) is a batch code where each data item is stored without change, i.e., each stored data item is a copy of one of the n data items. One of the basic yet challenging problems is to find optimal CBCs, i.e., CBCs for which total storage (N) is minimal for given values of n, m, k, and t. In [2], Paterson, Stinson and Wei exclusively studied CBCs and gave constructions of some optimal CBCs. In this article, we give a lower bound on the total storage (N) for CBCs. We give explicit construction of optimal CBCs for a range of values of n. For a different range of values of n, we give explicit construction of optimal and almost optimal CBCs. Our results partly settle an open problem of [2]. △ Less

Submitted 24 February, 2011; originally announced February 2011.

MSC Class: 05C65; 05D15 ACM Class: E.4; G.2.2

A note on the PT-invariant periodic potential V(x)=4 cos^2 x + 4 i V_0 sin 2x

Authors: Bikashkali Midya, Barnana Roy, Rajkumar Roychoudhury

Abstract: It is shown that the PT symmetric Hamiltonian with the periodic potential V(x) = 4 cos^2 x + 4 i V_0 sin 2x can be mapped into a Hermitian Hamiltonian for $V_0<0.5$, by a similarity transformation. It is also shown that there exist a second critical point of the potential V(x), apart from the known critical point $V_0=0.5$, for $V_0^c ~ .888437$ after which no part of the eigenvalues and the ban… ▽ More It is shown that the PT symmetric Hamiltonian with the periodic potential V(x) = 4 cos^2 x + 4 i V_0 sin 2x can be mapped into a Hermitian Hamiltonian for $V_0<0.5$, by a similarity transformation. It is also shown that there exist a second critical point of the potential V(x), apart from the known critical point $V_0=0.5$, for $V_0^c ~ .888437$ after which no part of the eigenvalues and the band structure remains real. Relevant physical consequence of this finding has been pointed out. △ Less

Submitted 19 April, 2010; originally announced April 2010.

Journal ref: Phys.Lett.A374:2605-2607,2010

Position Dependent Mass Schroedinger Equation and Isospectral Potentials : Intertwining Operator approach

Authors: Bikashkali Midya, Barnana Roy, Rajkumar Roychoudhury

Abstract: Here we have studied first and second-order intertwining approach to generate isospectral partner potentials of position-dependent (effective) mass Schroedinger equation. The second-order intertwiner is constructed directly by taking it as second order linear differential operator with position depndent coefficients and the system of equations arising from the intertwining relationship is solved… ▽ More Here we have studied first and second-order intertwining approach to generate isospectral partner potentials of position-dependent (effective) mass Schroedinger equation. The second-order intertwiner is constructed directly by taking it as second order linear differential operator with position depndent coefficients and the system of equations arising from the intertwining relationship is solved for the coefficients by taking an ansatz. A complete scheme for obtaining general solution is obtained which is valid for any arbitrary potential and mass function. The proposed technique allows us to generate isospectral potentials with the following spectral modifications: (i) to add new bound state(s), (ii) to remove bound state(s) and (iii) to leave the spectrum unaffected. To explain our findings with the help of an illustration, we have used point canonical transformation (PCT) to obtain the general solution of the position dependent mass Schrodinger equation corresponding to a potential and mass function. It is shown that our results are consistent with the formulation of type A N-fold supersymmetry [14,18] for the particular case N = 1 and N = 2 respectively. △ Less

Submitted 9 February, 2010; v1 submitted 7 January, 2010; originally announced January 2010.

Comments: Some references have been added

Journal ref: Journal of Mathematical Physics 51, 022109 (2010)

Strategic Execution in the Presence of an Uninformed Arbitrageur

Authors: Ciamac C. Moallemi, Beomsoo Park, Benjamin Van Roy

Abstract: We consider a trader who aims to liquidate a large position in the presence of an arbitrageur who hopes to profit from the trader's activity. The arbitrageur is uncertain about the trader's position and learns from observed price fluctuations. This is a dynamic game with asymmetric information. We present an algorithm for computing perfect Bayesian equilibrium behavior and conduct numerical expe… ▽ More We consider a trader who aims to liquidate a large position in the presence of an arbitrageur who hopes to profit from the trader's activity. The arbitrageur is uncertain about the trader's position and learns from observed price fluctuations. This is a dynamic game with asymmetric information. We present an algorithm for computing perfect Bayesian equilibrium behavior and conduct numerical experiments. Our results demonstrate that the trader's strategy differs significantly from one that would be optimal in the absence of the arbitrageur. In particular, the trader must balance the conflicting desires of minimizing price impact and minimizing information that is signaled through trading. Accounting for information signaling and the presence of strategic adversaries can greatly reduce execution costs. △ Less

Submitted 11 March, 2009; v1 submitted 18 January, 2008; originally announced January 2008.

Convergence of the Min-Sum Algorithm for Convex Optimization

Authors: Ciamac C. Moallemi, Benjamin Van Roy

Abstract: We establish that the min-sum message-passing algorithm and its asynchronous variants converge for a large class of unconstrained convex optimization problems. We establish that the min-sum message-passing algorithm and its asynchronous variants converge for a large class of unconstrained convex optimization problems. △ Less

Submitted 29 May, 2007; originally announced May 2007.