Searcharxiv⌕ Search

arXiv subjects

B. L. Wang

Publications and source records attributed to B. L. Wang.

6 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↗
arXiv2026

MiroThinker-1.7 & H1: Towards Heavy-Duty Research Agents via Verification

We present MiroThinker-1.7, a new research agent designed for complex long-horizon reasoning tasks. Building on this foundation, we further introduce MiroThinker-H1, which extends the agent with heavy-duty reasoning capabilities for more reliable multi-step problem solving. In particular, MiroThinker-1.7 improves the reliability of each interaction step through an agentic mid-training stage that emphasizes structured planning, contextual reasoning, and tool interaction. This enables more effective multi-step interaction and sustained reasoning across complex tasks. MiroThinker-H1 further incorporates verification directly into the reasoning process at both local and global levels. Intermediate reasoning decisions can be evaluated and refined during inference, while the overall reasoning trajectory is audited to ensure that final answers are supported by coherent chains of evidence. Across benchmarks covering open-web research, scientific reasoning, and financial analysis, MiroThinker-H1 achieves state-of-the-art performance on deep research tasks while maintaining strong results on specialized domains. We also release MiroThinker-1.7 and MiroThinker-1.7-mini as open-source models, providing competitive research-agent capabilities with significantly improved efficiency.

cs.CL↗
arXiv2022

Track-based alignment for the BESIII CGEM detector in the cosmic-ray test

The Beijing Electron Spectrometer III (BESIII) is a multipurpose detector operating on the Beijing Electron Positron Collider II (BEPCII). After more than ten year's operation, the efficiency of the inner layers of the Main Drift Chamber (MDC) decreased significantly. To solve this issue, the BESIII collaboration is planning to replace the inner part of the MDC with three layers of Cylindrical triple Gas Electron Multipliers (CGEM). The transverse plane spatial resolution of CGEM is required to be 120 $μ$m or better. To meet this goal, a careful calibration of the detector is necessary to fully exploit the potential of the CGEM detector. In all the calibrations, the detector alignment plays an important role to improve the detector precision. The track-based alignment for the CGEM detector with the Millepede algorithm is implemented to reduce the uncertainties of the hit position measurement. Using the cosmic-ray data taken in 2020 with the two layers setup, the displacement and rotation of the outer layer with respect to the inner layer is determined by a simultaneous fit applied to more than 160000 tracks. A good alignment precision has been achieved that guarantees the design request could be satisfied in the future. A further alignment is going to be performed using the combined information of tracks from cosmic-ray and collisions after the CGEM is installed into the BESIII detector.

hep-ex↗
arXiv2002

Weak UCP and perturbed monopole equations

We give a simple proof of weak Unique Continuation Property for perturbed Dirac operators, using the Carleman inequality. We apply the result to a class of perturbations of the Seiberg-Witten monopole equations that arise in Floer theory.

math.DG↗
arXiv1995

Topological Quantum Field Theory and Seiberg-Witten Monopoles

A topological quantum field theory is introduced which reproduces the Seiberg-Witten invariants of four-manifolds. Dimensional reduction of this topological field theory leads to a new one in three dimensions. Its partition function yields a three-manifold invariant, which can be regarded as the Seiberg-Witten version of Casson's invariant. A Geometrical interpretation of the three dimensional quantum field theory is also given.

hep-th↗