Search

We present the development of an automated scanning probe microscopy (SPM) measurement system using an advanced large-scale language model (LLM). This SPM system can receive instructions via social networking services (SNS), and the integration of SNS and LLMs enables real-time, language-agnostic control of SPM operations, thereby improving accessibility and efficiency. The integration of LLMs with AI systems with specialized functions brings the realization of self-driving labs closer.

We present an advanced scanning probe microscopy system enhanced with artificial intelligence (AI-SPM) designed for self-driving atomic-scale measurements. This system expertly identifies and manipulates atomic positions with high precision, autonomously performing tasks such as spectroscopic data acquisition and atomic adjustment. An outstanding feature of AI-SPM is its ability to detect and adapt to surface defects, targeting or avoiding them as necessary. It's also engineered to address typical challenges such as positional drift and tip apex atomic variations due to the thermal effect, ensuring accurate, site-specific surface analyses. Our tests under the demanding conditions of room temperature have demonstrated the robustness of the system, successfully navigating thermal drift and tip fluctuations. During these tests on the Si(111)-(7x7) surface, AI-SPM autonomously identified defect-free regions and performed a large number of current-voltage spectroscopy measurements at different adatom sites, while autonomously compensating for thermal drift and monitoring probe health. These experiments produce extensive data sets that are critical for reliable materials characterization and demonstrate the potential of AI-SPM to significantly improve data acquisition. The integration of AI into SPM technologies represents a step toward more effective, precise and reliable atomic-level surface analysis, revolutionizing materials characterization methods.

A machine-learning method for extracting the short-range part of the probe-surface interaction from force spectroscopy curves is presented. Our machine-learning algorithm consists of two stages: the first stage determines a boundary that separates the region where the short-range interaction is dominantly acting on the probe, and a second stage that finds the parameters to fit the interaction over the long-range region. We successfully applied this method to force spectroscopy maps acquired over the Si(111)-(7x7) surface and found, as a result, a faint structure on the short-range interaction for one of the probes used in the experiments that would have probably been obviated using human-supervised fitting strategies.

In this paper, we prove a combinatorial property of flows on a cycle. $C(V,E)$ is an undirected cycle with two commodities: $\{s_{1},t_{1}\}, \{s_{2},t_{2}\}$;$r_1>0,r_2>0, \mathbf r=(r_i)_{i=1,2}$ and $f,f'$ are both feasible flows for $(C,(s_i,t_i)_{i=1,2},\mathbf r)$. Then $\exists i\in\{1,2\}, p\in P_i, f(p)>0, \forall e\in p, f(e)\geq f'(e)$ ; Here for each $i\in\{1,2\}$, let $P_i$ be the set of $s_i$-$t_i$ paths in $C$ and $P=\cup_{i=1,2}P_i$. This means given a two-commodity instance on a cycle, any two distinct network flow $f$ and $f'$, compared with $f'$, $f$ can't decrease every path's flow amount at the same time. This combinatorial property is a generalization from single-commodity case to two-commodity case, and we also give an instance to illustrate the combinatorial property doesn't hold on for $k-$commodity case when $k\geq 3$.

Let $H=(V,E)$ be a hypergraph with vertex set $V$ and edge set $E$. $S\subseteq V$ is a feedback vertex set (FVS) of $H$ if $H\setminus S$ has no cycle and $\tau_c(H)$ denote the minimum cardinality of a FVS of $H$. In this paper, we prove $(i)$ if $H$ is a linear $3$-uniform hypergraph with $m$ edges, then $\tau_c(H)\le m/3$. $(ii)$ if $H$ is a $3$-uniform hypergraph with $m$ edges, then $\tau_c(H)\le m/2$ and furthermore, the equality holds on if and only if every component of $H$ is a $2-$cycle. Let $H=(V,E)$ be a hypergraph with vertex set $V$ and edge set $E$. $A\subseteq E$ is a feedback edge set (FES) of $H$ if $H\setminus A$ has no cycle and $\tau_c'(H)$ denote the minimum cardinality of a FES of $H$. In this paper, we prove if $H$ is a $3$-uniform hypergraph with $p$ components, then $\tau_c'(H)\le 2m-n+p$.

Given a weighted graph $G(V,E)$ with weight $\mathbf w: E\rightarrow Z^{|E|}_{+}$. A $k-$cycle covering is an edge subset $A$ of $E$ such that $G-A$ has no $k-$cycle. The minimum weight of $k-$cycle covering is the weighted covering number on $k-$cycle, denoted by $\tau_{k}(G_{w})$. In this paper, we design a $k-1/2$ approximation algorithm for the weighted covering number on $k-$cycle when $k$ is odd. Given a weighted graph $G(V,E)$ with weight $\mathbf w: E\rightarrow Z^{|E|}_{+}$. A $k-$clique covering is an edge subset $A$ of $E$ such that $G-A$ has no $k-$clique. The minimum weight of $k-$clique covering is the weighted covering number on $k-$clique, denoted by $\widetilde{\tau_{k}}(G_{w})$. In this paper, we design a $(k^{2}-k-1)/2$ approximation algorithm for the weighted covering number on $k-$clique. Last, we discuss the relationship between $k-$clique covering and $k-$clique packing in complete graph $K_{n}$.

Given a hypergraph H(V;E), a set of vertices S in V is a vertex cover if every edge has at least a vertex in S. The vertex cover number is the minimum cardinality of a vertex cover, denoted by t(H). In this paper, we prove that for every 3 uniform connected hypergraph H(V;E), t(H)<=(2m+1)/3 holds on where m is the number of edges. Furthermore, the equality holds on if and only if H(V;E) is a hypertree with perfect matching.

Given a simple graph $G=(V,E)$, a subset of $E$ is called a triangle cover if it intersects each triangle of $G$. Let $\nu_t(G)$ and $\tau_t(G)$ denote the maximum number of pairwise edge-disjoint triangles in $G$ and the minimum cardinality of a triangle cover of $G$, respectively. Tuza conjectured in 1981 that $\tau_t(G)/\nu_t(G)\le2$ holds for every graph $G$. In this paper, using a hypergraph approach, we design polynomial-time combinatorial algorithms for finding small triangle covers. These algorithms imply new sufficient conditions for Tuza's conjecture on covering and packing triangles. More precisely, suppose that the set $\mathscr T_G$ of triangles covers all edges in $G$. We show that a triangle cover of $G$ with cardinality at most $2\nu_t(G)$ can be found in polynomial time if one of the following conditions is satisfied: (i) $\nu_t(G)/|\mathscr T_G|\ge\frac13$, (ii) $\nu_t(G)/|E|\ge\frac14$, (iii) $|E|/|\mathscr T_G|\ge2$. Keywords: Triangle cover, Triangle packing, Linear 3-uniform hypergraphs, Combinatorial algorithms