Search

In this paper, we conduct a study to utilize LLMs as a solution for decision making that requires complex data analysis. We define Decision QA as the task of answering the best decision, $d_{best}$, for a decision-making question $Q$, business rules $R$ and a database $D$. Since there is no benchmark that can examine Decision QA, we propose Decision QA benchmark, DQA. It has two scenarios, Locating and Building, constructed from two video games (Europa Universalis IV and Victoria 3) that have almost the same goal as Decision QA. To address Decision QA effectively, we also propose a new RAG technique called the iterative plan-then-retrieval augmented generation (PlanRAG). Our PlanRAG-based LM generates the plan for decision making as the first step, and the retriever generates the queries for data analysis as the second step. The proposed method outperforms the state-of-the-art iterative RAG method by 15.8% in the Locating scenario and by 7.4% in the Building scenario, respectively. We release our code and benchmark at https://github.com/myeon9h/PlanRAG., Comment: NAACL 2024

In 1730, Euler defined the Gamma function $\Gamma(x)$ by the integral representation. It possesses many interesting properties and has wide applications in various branches of mathematics and sciences. According to Lerch, the Gamma function $\Gamma(x)$ can also be defined by the derivative of the Hurwitz zeta function $$\zeta(z,x)=\sum_{n=0}^{\infty}\frac{1}{(n+x)^{z}}$$ at $z=0$. Recently, Hu and Kim defined the corresponding Stieltjes constants $\widetilde{\gamma}_{k}(x)$ and Euler constant $\widetilde{\gamma}_{0}$ from the Taylor series of the alternating Hurwitz zeta function $\zeta_{E}(z,x)$ $$\zeta_{E}(z,x)=\sum_{n=0}^{\infty}\frac{(-1)^n}{(n+x)^z}.$$ And they also introduced the corresponding Gamma function $\widetilde{\Gamma}(x)$ which has the following Weierstrass--Hadamard type product $$\widetilde{\Gamma}(x)=\frac{1}{x}e^{\widetilde{\gamma}_{0}x}\prod_{k=1}^{\infty}\left(e^{-\frac{x}{k}}\left(1+\frac{x}{k}\right)\right)^{(-1)^{k+1}}.$$ In this paper, we shall further investigate the function $\widetilde{\Gamma}(x)$, that is, we obtain several properties in analogy to the classical Gamma function $\Gamma(x)$, including the integral representation, the limit representation, the recursive formula, the special values, the log-convexity, the duplication formula and the reflection equation. Furthermore, we also prove a Lerch-type formula, which shows that the derivative of $\zeta_{E}(z,x)$ can be representative by $\widetilde\Gamma(x)$. As an application to Stark's conjecture in algebraic number theory, we will explicit calculate the derivatives of the partial zeta functions for the maximal real subfield of cyclotomic fields at $z=0$., Comment: 24 pages

Let $\left\{E_n\right\}_{n=1}^{\infty}$ be the set of energy levels corresponding to a Hamiltonian $H$. Denote by $$\lambda_{0}=0~~\textrm{and}~~\lambda_{n}=E_{n}$$ for $n\in\mathbb N.$ In this paper, we shall construct and investigate the $p$-adic counterparts of the Hurwitz-type spectral zeta function \begin{equation} \zeta^{H}(s,\lambda)=\sum_{n=0}^{\infty}\frac{1}{(\lambda_{n}+\lambda)^{s}} \end{equation} and its alternating form \begin{equation} \zeta_{E}^{H}(s,\lambda)=2\sum_{n=0}^{\infty}\frac{(-1)^{n}}{(\lambda_{n}+\lambda)^{s}} \end{equation} in a parallel way., Comment: 20 pages

Scaling superconducting digital circuits requires fundamental changes in the current material set and fabrication process. The transition to 300 mm wafers and the implementation of advanced lithography are instrumental in facilitating mature CMOS processes, ensuring uniformity, and optimizing the yield. This study explores the properties of NbxTi(1-x)N films fabricated by magnetron DC sputtering on 300 mm Si wafers. As a promising alternative to traditional Nb in device manufacturing, NbxTi(1-x)N offers numerous advantages, including enhanced stability and scalability to smaller dimensions, in both processing and design. As a ternary material, NbxTi(1-x)N allows engineering material parameters by changing deposition conditions. The engineered properties can be used to modulate device parameters through the stack and mitigate failure modes. We report characterization of NbxTi(1-x)N films at less than 2% thickness variability, 2.4% Tc variability and 3% composition variability. The films material properties such as resistivity (140-375 {\Omega}cm) and critical temperature Tc (4.6 K - 14.1 K) are correlated with stoichiometry and morphology of the films. Our results highlight the significant influence of deposition conditions on crystallographic texture along the films and its correlation with Tc., Comment: 8 pages 8 figures

In his second notebook, Ramanujan discovered the following identity for the special values of $\zeta(s)$ at the odd positive integers \begin{equation*}\begin{aligned}\alpha^{-m}\,\left\{\dfrac{1}{2}\,\zeta(2m + 1) + \sum_{n = 1}^{\infty}\dfrac{n^{-2m - 1}}{e^{2\alpha n} - 1}\right\} &-(- \beta)^{-m}\,\left\{\dfrac{1}{2}\,\zeta(2m + 1) + \sum_{n = 1}^{\infty}\dfrac{n^{-2m - 1}}{e^{2\beta n} - 1}\right\}\nonumber &=2^{2m}\sum_{k = 0}^{m + 1}\dfrac{\left(-1\right)^{k-1}B_{2k}\,B_{2m - 2k+2}}{\left(2k\right)!\left(2m -2k+2\right)!}\,\alpha^{m - k + 1}\beta^k \label{(1.2)},\end{aligned} \end{equation*} where $ \alpha $ and $ \beta $ are positive numbers such that $ \alpha\beta = \pi^2 $ and $ m $ is a positive integer. As shown by Berndt in the viewpoint of general transformation of analytic Eisenstein series, it is a natural companion of Euler's famous formula for even zeta values. In this note, we prove an analogue of the above Ramanujan's identity in the functions fields setting, which involves the Bernoulli-Carlitz numbers., Comment: 8 pages

We describe a fabrication stackup for digital logic with 16 superconducting NbTiN layers, self-shunted a-silicon barrier Josephson Junctions (JJs), and low loss, high-$\kappa$ tunable HZO capacitors. The stack enables 400 MJJ/cm$^2$ device density, efficient routing, and AC power distribution on a resonant network. The materials scale beyond 28nm lithography and are compatible with standard high-temperature CMOS processes. We report initial results for two-metal layer NbTiN wires with 50nm critical dimension. A semi-ascendance wire-and-via process module using 193i lithography and 50nm critical dimension has shown cross-section uniformity of 1%=1s across the 300mm wafer, critical temperature of 12.5K, and critical current of 0.1mA at 4.2K. We also present a new design of the resonant AC power network enabled by NbTiN wires and HZO MIM capacitors. The design matches the device density and provides a 30 GHz clock with estimated efficiency of up to 90%. Finally, magnetic imaging of patterned NbTiN ground planes shows low intrinsic defectivity and consistent trapping of vorteces in 0.5 mm holes spaced on a 20 $\mu$m x 20 $\mu$m grid., Comment: 7 pages, 3 figures

Let $\{a_{1}, a_{2},\ldots, a_{n},\ldots\}$ be a sequence of complex numbers which has at most polynomial growth and satisfies an extra assumption. In this paper, inspired by a recent work of Sasane, we give an explanation of the sum $$a_{1}+2a_{2}+3a_{3}+\cdots+na_{n}+\cdots,$$ and more generally, for any $k\in\mathbb{N},$ the sum $$1^{k}a_{1}+2^{k}a_{2}+3^{k}a_{3}+\cdots+n^{k}a_{n}+\cdots,$$ from the viewpoint of distributions. As applications, we explain the following summation formulas \begin{equation*} \begin{aligned} 1^{k}-2^{k}+3^{k}-\cdots&=-\frac{E_{k}(0)}{2}, \\ 1^{k}+2^{k}+3^{k}+\cdots&=-\frac{B_{k+1}}{k+1}, \\ \epsilon^{1}1^{k}+\epsilon^{2}2^{k}+\epsilon^{3}3^{k}+\cdots&=-\frac{B_{k+1}(\epsilon)}{k+1}, \end{aligned} \end{equation*} where $E_{k}(0)$, $B_{k}$ and $B_{k}(\epsilon)$ are the Euler polynomials at 0, the Bernoulli numbers and the Apostol--Bernoulli numbers, respectively., Comment: 20 pages

In this note, we extend Euler's transformation formula from the alternating series to more general series. Then we give new expressions for the Riemann zeta function $\zeta(s)$ by the generalized difference operator $\Delta_{c}$, which provide analytic continuation of $\zeta(s)$ and new ways to evaluate the special values of $\zeta(-m)$ for $m=0,1,2,\ldots$. Applying these results, we further extend Huylebrouck's generalization of Wallis' well-known formula for $\pi$ in the half planes Re$(s)>0$ and Re$(s)>-1$, respectively. They imply several interesting special cases including $$ \frac{2\pi}{3^{\frac{3}{2}}}=\frac{3^{\frac{4}{3}}}{2^{\frac{4}{3}}} \frac{2^{\frac{1}{3}}\cdot3^{\frac{1}{3}}\cdot3^{\frac{1}{3}}\cdot4^{\frac{1}{3}}\cdot6^{\frac{2}{3}}\cdot6^{\frac{2}{3}}}{4^{\frac{1}{3}}\cdot4^{\frac{1}{3}}\cdot5^{\frac{1}{3}}\cdot5^{\frac{1}{3}}\cdot4^{\frac{2}{3}}\cdot5^{\frac{2}{3}}}\cdots, $$ $$ 3^{\gamma-\frac{\log 3}{2}}=\frac{3^{\frac{1}{3}}\cdot3^{\frac{1}{3}}}{2^{\frac{1}{2}}\cdot4^{\frac{1}{4}}} \frac{6^{\frac{1}{6}}\cdot6^{\frac{1}{6}}}{5^{\frac{1}{5}}\cdot7^{\frac{1}{7}}}\frac{9^{\frac{1}{9}}\cdot9^{\frac{1}{9}}}{8^{\frac{1}{8}}\cdot10^{\frac{1}{10}}}\cdots, $$ and $$ \left(3\left(\frac{2\pi e^{\gamma}}{A^{12}}\right)^{2}\right)^{\frac{\pi^2}{18}}=\frac{3^{\frac{1}{3^2}}\cdot3^{\frac{1}{3^2}}}{2^{\frac{1}{2^2}}\cdot4^{\frac{1}{4^2}}} \frac{6^{\frac{1}{6^2}}\cdot6^{\frac{1}{6^2}}}{5^{\frac{1}{5^2}}\cdot7^{\frac{1}{7^2}}}\frac{9^{\frac{1}{9^2}}\cdot9^{\frac{1}{9^2}}}{8^{\frac{1}{8^2}}\cdot10^{\frac{1}{10^2}}}\cdots,$$ where $\gamma$ is the Euler-Mascheroni constant and $A$ is the Glaisher-Kinkelin constant., Comment: 14 pages

In 1769, Euler proved the following result $$ \int_0^{\frac\pi2}\log(\sin \theta) d\theta=-\frac\pi2 \log2. $$ In this paper, as a generalization, we evaluate the definite integrals $$ \int_0^x \theta^{r-2}\log\left(\cos\frac\theta2\right)d\theta $$ for $r=2,3,4,\ldots.$ We show that it can be expressed by the special values of Kurokawa and Koyama's multiple cosine functions $\mathcal{C}_r(x)$ or by the special values of alternating zeta and Dirichlet lambda functions. In particular, we get the following explicit expression of the zeta value $$ \zeta(3)=\frac{4\pi^2}{21}\log\left(\frac{e^{\frac{4G}{\pi}}\mathcal{C}_3\left(\frac14\right)^{16}}{\sqrt2}\right), $$ where $G$ is Catalan's constant and $\mathcal{C}_3\left(\frac14\right)$ is the special value of Kurokawa and Koyama's multiple cosine function $\mathcal{C}_3(x)$ at $\frac14$. Furthermore, we prove several series representations for the logarithm of multiple cosine functions $\log\mathcal{C}_r\left(\frac x{2}\right)$ by zeta functions, $L$-functions or polylogarithms. One of them leads to another expression of $\zeta(3)$: $$\zeta(3)=\frac{72\pi^2}{11}\log\left(\frac{3^{\frac1{72}}\mathcal{C}_3\left(\frac16\right)}{\mathcal{C}_2\left(\frac16\right)^{\frac13}}\right).$$, Comment: 30 pages. We are grateful to Professor Jean-Paul Allouche for his interested in this paper and for his many helpful comments and suggestions

At the international congress of mathematicians in 1900, Hilbert claimed that the Riemann zeta function $\zeta(s)$ is not the solution of any algebraic ordinary differential equations on its region of analyticity. Let $T$ be an infinite order linear differential operator introduced by Van Gorder in 2015. Recently, Prado and Klinger-Logan (J. Number Theory 217 (2020), 422-442) showed that the Hurwitz zeta function $\zeta(s,a)$ formally satisfies the following linear differential equation $$T\left[\zeta (s,a) - \frac{1}{a^s}\right] = \frac{1}{(s-1)a^{s-1}}. $$ Then in (Abh. Math. Semin. Univ. Hambg. 91 (2021), no.1, 117-135), by defining $T_{p}^{a}$, a $p$-adic analogue of Van Gorder's operator $T,$ we constructed the following convergent infinite order linear differential equation satisfied by the $p$-adic Hurwitz-type Euler zeta function $\zeta_{p,E}(s,a)$ $$ T_{p}^{a}\left[\zeta_{p,E}(s,a)-\langle a\rangle^{1-s}\right] =\frac{1}{s-1}\left(\langle a-1 \rangle^{1-s}-\langle a\rangle^{1-s}\right). $$ In this paper, we consider this problem in the positive characteristic case. That is, by introducing $\zeta_{\infty}(s_{0},s,a,z)$, a Hurwitz type refinement of Goss zeta function, and an infinite order linear difference operator $L$, we establish the following difference equation \begin{equation*} L\left[\zeta_{\infty}\left(\frac{1}{T},s,a,0\right)\right]=\sum_{\alpha\in\mathbb{F}_{q}} \frac{1}{\langle a+\alpha\rangle^{s}}. \end{equation*}, Comment: 14 page

Let $\tilde\Gamma(z)$ be the modified gamma function introduced by the authors in a recent preprint "arXiv2106.14674". In this note, we obtain the following Mizuno-type result: \begin{equation*} \prod_{m=0}^{\infty}\left\{\prod_{j=1}^{n}(m+z_{j})\right\}^{(-1)^{m}}=\frac{\left(\sqrt{\frac{\pi}{2}}\right)^n}{\prod_{j=1}^{n}\tilde\Gamma(z_{j})}, \end{equation*} which imply a Kurokawa--Wakayama type formula \begin{equation*} \prod_{m=0}^\infty\left((m+x)^{n}-y^n\right)^{(-1)^{m}} =\frac{\left(\sqrt{\frac{\pi}{2}}\right)^n}{\prod_{\zeta^{n}=1}\tilde\Gamma(x-\zeta y)} \end{equation*} and a Lerch-type formula \begin{equation*} \prod_{m=0}^\infty(m+x)^{(-1)^{m}}=\frac{\sqrt{\frac{\pi}{2}}}{\tilde\Gamma(x)}. \end{equation*} By setting $x=1$ in the above result, we recover Wallis' 1656 fomula \begin{equation*}\frac{2\cdot2}{1\cdot 3}\frac{4\cdot4}{3\cdot 5}\frac{6\cdot6}{5\cdot 7}\cdots=\frac{\pi}{2}. \end{equation*}, Comment: 11 pages

The classical Lerch's formula states the following normalized product: $$\prod_{n=0}^\infty(x+n)=\frac{\sqrt{2\pi}}{\Gamma(x)},\quad \textrm{Re}(x)>0, $$ where $\Gamma(x)$ is the Euler gamma function. In this note, by using Barnes' multiple zeta function and its alternating form, we obtain two kinds of generalizations of Lerch's formula, which imply the product $$ \prod_{n=1}^\infty n=\sqrt{2\pi} $$ (in the sense of zeta regularization) and the product $$\frac{2\cdot2}{1\cdot 3}\frac{4\cdot4}{3\cdot 5}\frac{6\cdot6}{5\cdot 7}\cdots=\frac{\pi}{2}$$ (Wallis' formula in 1656), respectively., Comment: 10 pages

Dating back to Euler, in classical analysis and number theory, the Hurwitz zeta function $$ \zeta(z,q)=\sum_{n=0}^{\infty}\frac{1}{(n+q)^{z}}, $$ the Riemann zeta function $\zeta(z)$, the generalized Stieltjes constants $\gamma_k(q)$, the Euler constant $\gamma$, Euler's gamma function $\Gamma(q)$ and the digamma function $\psi(q)$ have many close connections on their definitions and properties. There are also many integrals, series or infinite product representations of them along the history. In this note, we try to provide a parallel story for the alternating Hurwitz zeta function (also known as the Hurwitz-type Euler zeta function) $$\zeta_{E}(z,q)=\sum_{n=0}^\infty\frac{(-1)^{n}}{(n+q)^{z}},$$ the alternating zeta function $\zeta_{E}(z)$ (also known as the Dirichlet's eta function $\eta(z)$), the modified Stieltjes constants $\tilde\gamma_k(q)$, the modified Euler constant $\tilde\gamma_{0}$, the modified gamma function $\tilde\Gamma(q)$ and the modified digamma function $\tilde\psi(q)$ (also known as the Nielsen's $\beta$ function). Many new integrals, series or infinite product representations of these constants and special functions have been found. By the way, we also get two new series expansions of $\pi:$ \begin{equation*} \frac{\pi^2}{12}=\frac34-\sum_{k=1}^\infty(\zeta_E(2k+2)-1) \end{equation*} and \begin{equation*} \frac{\pi}{2}= \log2+2\sum_{k=1}^\infty\frac{(-1)^k}{k!}\tilde\gamma_k(1)\sum_{j=0}^kS(k,j)j!. \end{equation*}, Comment: 34 pages, Final version

We present a novel deep neural model for text detection in document images. For robust text detection in noisy scanned documents, the advantages of multi-task learning are adopted by adding an auxiliary task of text enhancement. Namely, our proposed model is designed to perform noise reduction and text region enhancement as well as text detection. Moreover, we enrich the training data for the model with synthesized document images that are fully labeled for text detection and enhancement, thus overcome the insufficiency of labeled document image data. For the effective exploitation of the synthetic and real data, the training process is separated in two phases. The first phase is training only synthetic data in a fully-supervised manner. Then real data with only detection labels are added in the second phase. The enhancement task for the real data is weakly-supervised with information from their detection labels. Our methods are demonstrated in a real document dataset with performances exceeding those of other text detection methods. Moreover, ablations are conducted and the results confirm the effectiveness of the synthetic data, auxiliary task, and weak-supervision. Whereas the existing text detection studies mostly focus on the text in scenes, our proposed method is optimized to the applications for the text in scanned documents.

By using Cauchy's formula, it is known that Bernoulli numbers and Euler numbers can be represented by the contour integrals \begin{equation*} \begin{aligned} B_n&=\frac{n!}{2\pi i}\oint \frac{z}{e^z-1}\frac{d z}{z^{n+1}},\label{condefi}\\[4pt] E_n&=\frac{n!}{2\pi i}\oint \frac{2e^z}{e^{2z}+1}\frac{d z}{z^{n+1}}, \end{aligned} \end{equation*} while the following Witt's formula represents Euler polynomials through the fermionic $p$-adic integrals $$E_n(a)=\int_{\mathbb Z_p}(x+a)^nd\mu_{-1}(x).$$ Base on the above Witt's identity and the binomial theorem, we prove some new identities for the Euler polynomials briefly. In particular, some symmetry properties of Euler polynomials have been discovered, which implies many interesting identities (known or unknown), including the Kaneko-Momiyama type identities (shown by Wu, Sun, and Pan) and the Alzer-Kwong type identity for Euler polynomials., Comment: 15 pages

In this paper, we introduce the concept of the (higher order) Appell-Carlitz numbers which unifies the definitions of several special numbers in positive characteristic, such as the Bernoulli-Carlitz numbers and the Cauchy-Carlitz numbers.Their generating function is usually named Hurwitz series in the function field arithmetic. By using Hasse-Teichm\"uller derivatives, we also obtain several properties of the (higher order) Appell-Carlitz numbers, including a recurrence formula, two closed forms expressions, and a determinant expression. The recurrence formula implies Carlitz's recurrence formula for Bernoulli-Carlitz numbers. Two closed from expressions implies the corresponding results for Bernoulli-Carlitz and Cauchy-Carlitz numbers . The determinant expression implies the corresponding results for Bernoulli-Carlitz and Cauchy-Carlitz numbers, which are analogues of the classical determinant expressions of Bernoulli and Cauchy numbers stated in an article by Glaisher in 1875., Comment: Final version; 16 pages