SearcharxivSearch

arXiv subjects

Thomas Hille

Publications and source records attributed to Thomas Hille.

5 recordsLinked to original sources

On the $σ$-Pair Correlation Density of Quadratic Sequences Modulo One

In this note we study the $σ$-pair correlation density \begin{equation*}R_2^σ([a,b], \{ θ_n \}_n, N)= \frac{1}{N^{2-σ}} \# \big \{ 1 \leq j \neq k \leq N \, \big| \, θ_{j} - θ_{k} \in \big [ \frac{a}{N^σ},\frac{b}{N^σ} \big ]+ \mathbb Z \big \} \end{equation*} of a sequence $\{ θ_n\}_n$ that is equidistributed modulo one for $0 \leq σ<2$. The case $σ=1$ is commonly referred to as the pair correlation density and the sequence $\{ n^2 α\}_n$ has been of special interest due to its connection to a conjecture of Berry and Tabor on the energy levels of generic completely integrable systems. We prove that if $α$ is Diophantine of type $3-ε$ for every $ε>0$, then for any $0 \leq σ<1$ \begin{align*} \mathrm R_2^σ([a,b], \{ αn^2 \}_n, N) \to b-a, \text{ as } N \to \infty. \end{align*} In this case, we say that the sequence exhibits $σ$-pair correlation. In addition to this, we show that for any $0 \leq σ< \frac{1}{4}(9 -\sqrt{17})=1.21922...$ there is a set of full Lebesgue measure such that the sequence $\{ αn^2 \}_n$ exhibits $σ$-pair correlation.

math.NT

Distribution of Values of Quadratic Forms at Integral Points

The number of lattice points in $d$-dimensional hyperbolic or elliptic shells $\{m : a<Q[m]<b\}$, which are restricted to rescaled and growing domains $r\;Ω$, is approximated by the volume. An effective error bound of order $o(r^{d-2})$ for this approximation is proved based on Diophantine approximation properties of the quadratic form $Q$. These results allow to show effective variants of previous non-effective results in the quantitative Oppenheim problem and extend known effective results in dimension $d \geq 9$ to dimension $d \geq 5$. They apply to wide shells when $b-a$ is growing with $r$ and to positive definite forms $Q$. For indefinite forms they provide explicit bounds (depending on the signature or Diophantine properties of $Q$) for the size of non-zero integral points $m$ in dimension $d\geq 5$ solving the Diophantine inequality $|Q[m]| < \varepsilon$ and provide error bounds comparable with those for positive forms up to powers of $\log r$.

math.NT

Small Values of Indefinite Diagonal Quadratic Forms at Integer Points in at least five Variables

For any $\varepsilon > 0$ we derive effective estimates for the size of a non-zero integral point $m \in \mathbb{Z}^d \setminus \{0\}$ solving the Diophantine inequality $\lvert Q[m] \rvert < \varepsilon$, where $Q[m] = q_1 m_1^2 + \ldots + q_d m_d^2$ denotes a non-singular indefinite diagonal quadratic form in $d \geq 5$ variables. In order to prove our quantitative variant of the Oppenheim conjecture, we extend an approach developed by Birch and Davenport [BD58b] to higher dimensions combined with a theorem of Schlickewei [Sch85]. The result obtained is an optimal extension of Schlickewei's result, giving bounds on small zeros of integral quadratic forms depending on the signature $(r,s)$, to diagonal forms up to a negligible growth factor.

math.NT

Aspects on Finding the Optimal Practical Programming Exercise for MOOCs

Massive Open Online Courses (MOOCs) focus on manifold subjects, ranging from social sciences over languages to technical skills, and use different means to train the respective skills. MOOCs that are teaching programming skills aim to incorporate practical exercises into the course corpus to give students the hands-on experience necessary for understanding and mastering programming. These exercises, apart from technical challenges, come with a series of questions to be addressed, for example: which fraction of the participants' time should they take (compared to video lectures and other course activities), which difficulty should be aimed for, how much guidance should be offered and how much repetition should be incorporated? The perceived difficulty of a task depends on previous knowledge, supplied hints, the required time for solving and the number of failed attempts the participant made. Furthermore, the detail and accuracy of the problem description, the restrictiveness of the applied test cases and the preparation provided specifically for a given exercise also influence the perceived difficulty of a task. In this paper, we explore the data of three programming courses to find criteria for optimal practical programming exercises. Based on over 3 million executions and scoring runs of participants' task submissions, we aim to deduct exercise difficulty, student patterns in approaching the tasks and potential flaws in task descriptions and preparatory videos. We compare our findings to in class trainings and traditional, mostly video and quiz based MOOCs. Finally, we propose approaches and methods to improve programming courses for participants as well as instructors.

cs.SE

Effects of Automated Interventions in Programming Assignments: Evidence from a Field Experiment

A typical problem in MOOCs is the missing opportunity for course conductors to individually support students in overcoming their problems and misconceptions. This paper presents the results of automatically intervening on struggling students during programming exercises and offering peer feedback and tailored bonus exercises. To improve learning success, we do not want to abolish instructionally desired trial and error but reduce extensive struggle and demotivation. Therefore, we developed adaptive automatic just-in-time interventions to encourage students to ask for help if they require considerably more than average working time to solve an exercise. Additionally, we offered students bonus exercises tailored for their individual weaknesses. The approach was evaluated within a live course with over 5,000 active students via a survey and metrics gathered alongside. Results show that we can increase the call outs for help by up to 66% and lower the dwelling time until issuing action. Learnings from the experiments can further be used to pinpoint course material to be improved and tailor content to be audience specific.

cs.CY