arXiv subjects

Yikang Yang

Publications and source records attributed to Yikang Yang.

9 recordsLinked to original sources

arXiv2026

Procedura: Agentic 3D Modeling with Procedural Control

Native 3D generators now recover impressive mesh geometry from a single image. However, a dense mesh stays soft where a machined object should be sharp, it carries no part decomposition, and it exposes no parameter a user could edit. To address this, we explore the paradigm of 3D shape as code, leveraging and scaling the coding ability of an LLM for 3D modeling. We introduce Procedura, a novel 3D modeling agent framework that writes an object as a procedural assembly, a parametric program whose named parts are joined by typed, machine-checkable mates. From a text prompt, the agent plans the object as an assembly graph and writes the program part by part, solving each placement from the mated frames rather than guessing it, and admitting a part only once compile, mate, and connectivity checks pass. A decoupled vision critic then refines the assembly one diagnosed fix at a time. Moreover, the same graph carries per-part materials and a simulator-validated articulation. We evaluate on P3D-Bench under its assembly judge, and with the same judge on MechBench-36, our hard-surface benchmark. On both, Procedura outperforms state-of-the-art native 3D generators and every prior 3D-code agent on judged quality, produces the sharpest edges of any method we evaluate, and is the only one whose output is an editable, part-structured program.

cs.CV↗

arXiv2026

When AI Designs AI: Innovation or Imitation?

Recent advances in LLM agents have made them increasingly capable of designing methods for complex AI tasks. This raises two central questions about agent-designed methods relative to human-designed methods: how well they perform, and how different their algorithmic designs are. To study these questions, this paper introduces an analysis that derives task-specific algorithmic design spaces from human-designed methods, maps both human- and agent-designed methods into these spaces, and quantifies their algorithmic differences at the module level. Widely used LLM agents are evaluated on a suite of representative, open-ended AI tasks spanning multiple modalities, and the methods they design are analyzed in terms of both task performance and algorithmic differences from human-designed methods. Experimental results show that current agents can occasionally match or surpass human state-of-the-art (SOTA) performance (10/72 configurations), but such success does not generalize reliably across tasks or agents. Moreover, 96.8% of agent-designed methods fall within human-derived algorithmic design spaces, largely recombining algorithmic choices found in human-designed methods, while nearly half exactly match an existing human algorithmic design. Taken together, these findings suggest that although current agents can occasionally match or surpass human SOTA performance, their algorithmic designs remain within human-derived algorithmic design spaces, reflecting the reuse and recombination of algorithmic choices.

cs.AI↗

arXiv2026

P3D-Bench: Benchmarking MLLMs for Parametric 3D Generation and Structural Reasoning

Multimodal large language models can write code to produce complex programs as well as use programs to do 3D modeling, which opens up a new avenue for 3D generation powered by their priors, world knowledge and reasoning. Yet existing benchmarks rarely evaluate 3D modeling through code. Such modeling demands more than runnable code: from a text or visual specification, a model must generate a parametric 3D program that is geometrically precise, semantically aligned and assembly-consistent. We introduce P3D-Bench, a benchmark for parametric 3D generation. Unlike a 3D mesh, a parametric 3D program exposes explicit dimensions, construction operations and part relations, revealing whether a model recovers a design's structure, not just its appearance. Under a unified protocol, P3D-Bench covers three task families (Text-to-3D, Image-to-3D and Assembly-3D) and scores each output for executability, geometric fidelity, topology, text-grounded constraints, multiview semantic alignment and part-level structure. We evaluate frontier MLLMs and text-only LLMs on 400 text cases, 400 image cases and 203 annotated assemblies, with domain-specific models as reference points. Our extensive evaluation yields three findings. First, assemblies are the hardest setting, where models still fail to compose multiple parts into a coherent structure. Second, models can often recover the global shape and semantic identity of the target object, yet fail to reproduce the precise parametric geometry specified by the input. Third, part-level modeling remains weak on assemblies, where models recover neither the geometry of each part nor the right number of parts. These results position P3D-Bench as a benchmark for evaluating precise parametric geometry and part-level structure in parametric 3D generation.

cs.CV↗

arXiv2026

GraDE: A Graph Diffusion Estimator for Frequent Subgraph Discovery in Neural Architectures

Finding frequently occurring subgraph patterns or network motifs in neural architectures is crucial for optimizing efficiency, accelerating design, and uncovering structural insights. However, as the subgraph size increases, enumeration-based methods are perfectly accurate but computationally prohibitive, while sampling-based methods are computationally tractable but suffer from a severe decline in discovery capability. To address these challenges, this paper proposes GraDE, a diffusion-guided search framework that ensures both computational feasibility and discovery capability. The key innovation is the Graph Diffusion Estimator (GraDE), which is the first to introduce graph diffusion models to identify frequent subgraphs by scoring their typicality within the learned distribution. Comprehensive experiments demonstrate that the estimator achieves superior ranking accuracy, with up to 114\% improvement compared to sampling-based baselines. Benefiting from this, the proposed framework successfully discovers large-scale frequent patterns, achieving up to 30$\times$ higher median frequency than sampling-based methods.

cs.LG↗

arXiv2026

Probing Memes in LLMs: A Paradigm for the Entangled Evaluation World

Current evaluation paradigms for large language models (LLMs) characterize models and datasets separately, yielding coarse descriptions: items in datasets are treated as pre-labeled entries, and models are summarized by overall scores such as accuracy, together ignoring the diversity of population-level model behaviors across items with varying properties. To address this gap, this paper conceptualizes LLMs as composed of memes, a notion introduced by Dawkins as cultural genes that replicate knowledge and behavior. Building on this perspective, the Probing Memes paradigm reconceptualizes evaluation as an entangled world of models and data. It centers on a Perception Matrix that captures model-item interactions, enabling Probe Properties for characterizing items and Meme Scores for depicting model behavioral traits. Applied to 9 datasets and 4,507 LLMs, Probing Memes reveals hidden capability structures and quantifies phenomena invisible under traditional paradigms (e.g., elite models failing on problems that most models answer easily). It not only supports more informative and extensible benchmarks but also enables population-based evaluation of LLMs.

cs.CL↗

arXiv2025

Direct3D-S2: Gigascale 3D Generation Made Easy with Spatial Sparse Attention

Generating high-resolution 3D shapes using volumetric representations such as Signed Distance Functions (SDFs) presents substantial computational and memory challenges. We introduce Direct3D-S2, a scalable 3D generation framework based on sparse volumes that achieves superior output quality with dramatically reduced training costs. Our key innovation is the Spatial Sparse Attention (SSA) mechanism, which greatly enhances the efficiency of Diffusion Transformer (DiT) computations on sparse volumetric data. SSA allows the model to effectively process large token sets within sparse volumes, substantially reducing computational overhead and achieving a 3.9x speedup in the forward pass and a 9.6x speedup in the backward pass. Our framework also includes a variational autoencoder (VAE) that maintains a consistent sparse volumetric format across input, latent, and output stages. Compared to previous methods with heterogeneous representations in 3D VAE, this unified design significantly improves training efficiency and stability. Our model is trained on public available datasets, and experiments demonstrate that Direct3D-S2 not only surpasses state-of-the-art methods in generation quality and efficiency, but also enables training at 1024 resolution using only 8 GPUs, a task typically requiring at least 32 GPUs for volumetric representations at 256 resolution, thus making gigascale 3D generation both practical and accessible. Project page: https://www.neural4d.com/research/direct3d-s2.

cs.CV↗

arXiv2024

Hybrid Cooperative Co-Evolution Algorithm for Deadlock-prone Distributed Assembly Flowshop Scheduling with Limited buffers Using Petri nets

The distributed assembly flowshop scheduling problem (DAFSP) can be applied to immense manufacturing environments. In DAFSP, jobs are first processed in distributed flowshops, and then assembled into final products by an assembly machine, which usually has limited buffers in practical application. This limited capacity can lead to deadlocks, halting job completion and blocking the entire manufacturing process. However, existing scheduling methods fail to address these deadlocks in DAFSP effectively. As such, we develop a hybrid cooperative co-evolution (HCCE) algorithm for solving the deadlock-prone DAFSP by minimizing the makespan. For the first time, we use Petri nets to analyze the deadlocks in DAFSP and propose a Petri net-based deadlock amending method (IDAM), which is further integrated into HCCE to ensure the feasibility (i.e., deadlock-freeness) of solutions. Importantly, HCCE contains an elite archive (EAR) and two subpopulations. It uses the problem-specific operators for heuristic initialization and global-search. To enhance the quality and diversity of solutions, an information transfer mechanism (ITM) is developed among subpopulation and EAR, and four local-search operators are performed sequentially on each individual in EAR. Finally, comprehensive experiments demonstrate the effectiveness and superiority of the proposed HCCE algorithm.

cs.DC↗

arXiv2024

Achieving Consistent and Comparable CPU Evaluation

The challenge of CPU evaluation lies in the fact that user-perceived performance metrics can only be measured on an independently running system consisting of the CPU and other indispensable components, and hence it is difficult to accurately attribute the deviations in the evaluation outcomes to the differences between the CPUs. Our experiments reveal that the industry-standard CPU benchmark, SPEC CPU2017, suffers from a significant flaw: for the identical CPU, undefined configurations of other indispensable components introduce uncontrolled variability in evaluation outcomes. We propose a rigorous CPU evaluation methodology. Through theoretical analysis and pioneering controlled experiments, we systematically compare our methodology against four established methodologies: the SPEC CPU 2017, two DOE variants, and one RCTs approach. The results show our methodology can achieve consistent and comparable evaluation outcomes, while others exhibit inherent limations.

cs.PF↗

arXiv2023

A Linear Combination-based Method to Construct Proxy Benchmarks for Big Data Workloads

During early stages of CPU design, benchmarks can only run on simulators to evaluate CPU performance. However, most big data benchmarks are too huge at code size scale, which causes them to be unable to finish running on simulators at an acceptable time cost. Moreover, big data benchmarks usually need complex software stacks to support their running, which is hard to be ported on simulators. Proxy benchmarks, without long running times and complex software stacks, have the same micro-architectural metrics as real benchmarks, which means they can represent real benchmarks' micro-architectural characteristics. Therefore, proxy benchmarks can replace real benchmarks to run on simulators to evaluate the CPU performance. The biggest challenge is how to guarantee that the proxy benchmarks have exactly the same micro-architectural metrics as real benchmarks when the number of micro-architectural metrics is very large. To deal with this challenge, we propose a linear combination-based proxy benchmark generation methodology that transforms this problem into solving a linear equation system. We also design the corresponding algorithms to ensure the linear equation is astringency, which means that although sometimes the linear equation system doesn't have a unique solution, the algorithm can find the best solution by the non-negative least square method. We generate fifteen proxy benchmarks and evaluate their running time and accuracy in comparison to corresponding real benchmarks for Mysql and RockDB. On the typical Intel Xeon platform, the average running time is 1.62s, and the average accuracy of every micro-architectural metric is over 92%, while the longest running time of real benchmarks is nearly 4 hours. We also conduct two case studies that demonstrate that our proxy benchmarks are consistent with real benchmarks both before and after prefetch or Hyper-Threading is turned on.

cs.PF↗