arXiv subjects

S. Nie

Publications and source records attributed to S. Nie.

4 recordsLinked to original sources

arXiv2026

Apodex 1.1: Scaling Agentic Intelligence for Complex Work

General-purpose language models can reason and synthesize knowledge, but complex work also requires sustained interaction with files, information sources, and executable code, together with state maintenance, failure recovery, and verifiable delivery. We call this \emph{working capability}: sustained, verifiable progress toward a real-world objective. Apodex 1.1 develops this capability along two complementary dimensions. \emph{Environment Scaling} expands the diversity and verifiability of executable file, search, and code environments, while \emph{Agentic Coordination Scaling} trains agents to decompose long-horizon tasks, delegate parallel work, integrate asynchronous results, and replan. A shared execution harness and AgentOS maintain task state and provenance across tools and agents, and training turns environment trajectories and coordination traces into reliable behavior. Across complex professional work, finance, scientific research, mathematics, coding, and search, Apodex 1.1 reaches the leading performance band despite using a substantially smaller model than many frontier systems. The 35B-parameter Apodex 1.1 Mini further retains strong working capability in a locally deployable form. These results ground agentic intelligence in useful, verifiable work completed over time and advance our goal of building a \emph{Heavy-Duty Solver} for ambitious, long-running tasks.

cs.AI↗

arXiv2012

Growth Structure and Work Function of Bilayer Graphene on Pd(111)

Using in situ low-energy electron microscopy and density functional theory, we studied the growth structure and work function of bilayer graphene on Pd(111). Low-energy electron diffraction analysis established that the two graphene layers have multiple rotational orientations relative to each other and the substrate plane. We observed heterogeneous nucleation and simultaneous growth of multiple, faceted layers prior to the completion of second layer. We propose that the facetted shapes are due to the zigzag-terminated edges bounding graphene layers growing under the larger overlying layers. We also found that the work functions of bilayer graphene domains are higher than those of monolayer graphene, and depend sensitively on the orientations of both layers with respect to the substrate. Based on first-principles simulations, we attribute this behavior to oppositely oriented electrostatic dipoles at the graphene/Pd and graphene/graphene interfaces, whose strengths depend on the orientations of the two graphene layers.

cond-mat.mtrl-sci↗

arXiv2011

Origin of the Mosaicity in Graphene Grown on Cu(111)

We use low-energy electron microscopy to investigate how graphene grows on Cu(111). Graphene islands first nucleate at substrate defects such as step bunches and impurities. A considerable fraction of these islands can be rotationally misaligned with the substrate, generating grain boundaries upon interisland impingement. New rotational boundaries are also generated as graphene grows across substrate step bunches. Thus, rougher substrates lead to higher degrees of mosaicity than do flatter substrates. Increasing the growth temperature improves crystallographic alignment. We demonstrate that graphene growth on Cu(111) is surface diffusion limited by comparing simulations of the time evolution of island shapes with experiments. Islands are dendritic with distinct lobes, but unlike the polycrystalline, four-lobed islands observed on (100)-textured Cu foils, each island can be a single crystal. Thus, epitaxial graphene on smooth, clean Cu(111) has fewer structural defects than it does on Cu(100).

cond-mat.mtrl-sci↗

arXiv2010

Thickness monitoring of graphene on SiC using low-energy electron diffraction

The formation of epitaxial graphene on SiC is monitored in-situ using low-energy electron diffraction (LEED). The possibility of using LEED as an in-situ thickness monitor of the graphene is examined. The ratio of primary diffraction spot intensities for graphene compared to SiC is measured for a series of samples of known graphene thickness (determined using low-energy electron microscopy). It is found that this ratio is effective for determining graphene thicknesses in the range 1 to 3 monolayers. Effects of a distribution of graphene thicknesses on this method of thickness determination are considered.

cond-mat.mtrl-sci↗