arXiv subjects

Ruilin Chen

Publications and source records attributed to Ruilin Chen.

8 recordsLinked to original sources

arXiv2026

Representation-Aligned Tactile Grounding for Contact-Rich Robotic Manipulation

Tactile-enhanced vision-language-action (VLA) policies have been introduced for contact-rich manipulation, where critical interaction states are often hidden from vision. Future tactile prediction is a promising way to use touch because it turns tactile outcomes into supervision for action-induced contact dynamics. Yet VLA policies contain representations with different roles, from perceptual encoding to motor prediction, making it unclear where this supervision should be applied. We study this as a representation-alignment problem. Through a linear probe analysis, we find that future tactile states are most predictable from intermediate action-expert features, rather than from vision-language features or final action states. Motivated by this observation, we introduce a lightweight Latent Tactile Predictor (LTP), which predicts compact future tactile embeddings from the identified intermediate representation. By avoiding direct prediction of noisy raw tactile signals, LTP provides an action-outcome grounding signal that aligns intermediate action representations with future contact consequences. Experiments on real-world contact-rich manipulation tasks show that representation-aligned tactile grounding outperforms less aligned or multi-interface tactile prediction, highlighting the importance of where tactile supervision is applied.

cs.RO↗

arXiv2026

Optimus: A Generic Operator-Level PyTorch Model Transformation Framework

In large-scale industrial applications, deep learning models that power recommendation and ranking have complex and diverse model architectures. These models are continuously developed and refined by large teams of machine learning engineers, rendering manual optimization infeasible. Consequently, graph-based optimization techniques have become an industry standard for boosting performance, with PyTorch FX transformations leading the charge. These transformations typically rely on a set of human-engineered module-level rewrite rules which are not scalable to diverse model architectures. To address this limitation, we introduce Optimus, a general-purpose model transformation framework built in the PyTorch 2.x (PT2) machine learning compiler. With a concise set of predefined patterns, Optimus applies an efficient greedy search algorithm for pattern matching and replacement, while preserving model semantic. It is designed and implemented as a highly customizable and extensible framework integrated into the PT2 stack. Our evaluation shows that the framework can achieve up to 63% speedup, 6% peak memory reduction, and over 400 second compile time decrease for our industry-scale recommendation models compared to baselines. Optimus is open-sourced together with PyTorch 2.x as a customizable model transformation layer.

cs.PF↗

arXiv2025

A Technical Report on the Second Place Solution for the CIKM 2025 AnalytiCup Competition

In this work, we address the challenge of multilingual category relevance judgment in e-commerce search, where traditional ensemble-based systems improve accuracy but at the cost of heavy training, inference, and maintenance complexity. To overcome this limitation, we propose a simplified yet effective framework that leverages prompt engineering with Chain-of-Thought task decomposition to guide reasoning within a single large language model. Specifically, our approach decomposes the relevance judgment process into four interpretable subtasks: translation, intent understanding, category matching, and relevance judgment -- and fine-tunes a base model (Qwen2.5-14B) using Low-Rank Adaptation (LoRA) for efficient adaptation. This design not only reduces computational and storage overhead but also enhances interpretability by explicitly structuring the model's reasoning path. Experimental results show that our single-model framework achieves competitive accuracy and high inference efficiency, processing 20 samples per second on a single A100 GPU. In the CIKM 2025 AnalytiCup Competition Proposals, our method achieved 0.8902 on the public leaderboard and 0.8889 on the private leaderboard, validating the effectiveness and robustness of the proposed approach. These results highlight that structured prompting combined with lightweight fine-tuning can outperform complex ensemble systems, offering a new paradigm for scalable industrial AI applications.

cs.IR↗

arXiv2025

Output-recurrent gated state space model for multiphase flows modeling and uncertainty quantification of exhaust vehicles

This paper presents an Output-Recurrent Gated State Space Model (OR-GSSM) for complex multiphase flows modeling and uncertainty quantification of exhaust vehicles during motion. By establishing the state-space formulation of the gas-liquid Navier-Stokes equations applying semigroup theory and Galerkin projection, explicitly characterizing the dynamic coupling evolution between the velocity, pressure, and volume fraction fields. A novel Gated State Space Transition (GSST) unit is designed to learn parameterized transition and input matrices with adaptive timescales, enhancing physical interpretability and computational efficiency. The output recursion mechanism aligns with the numerical solution characteristics of state-space equations, mitigating long-term error accumulation and addressing training-inference pattern mismatch issues inherent in teacher forcing and scheduled sampling. Validations on the underwater cone-head and water-exit hemisphere-head vehicles demonstrate that: OR-GSSM outperforms OR-ConvLSTM and OR-ConvGRU baselines in accuracy and computational efficiency through its physics-informed adaptive state-space unit design and parallel matrix operations; The output recursion mechanism ensures more stable training, better generalization, and higher prediction accuracy than teacher forcing and scheduled sampling; OR-GSSM accurately captures the gas-phase expansion, gas-liquid mixing formation, backflow jet generation, bubble shedding, and entire water-exit process, etc, showcasing outstanding modeling capability; Its uncertainty quantification effectively characterizes flow features and uncertainty distributions, validating prediction reliability. The proposed method resolves the accuracy-real-time trade-off in traditional computational fluid dynamics, advancing machine learning for multiphase flow modeling and uncertainty quantification in exhaust vehicles.

physics.flu-dyn↗

arXiv2025

High-efficient machine learning projection method for incompressible Navier-Stokes equations

This study proposes a high-efficient machine learning (ML) projection method using forward-generated data for incompressible Navier-Stokes equations. A Poisson neural network (Poisson-NN) embedded method and a wavelet transform convolutional neural network multigrid (WTCNN-MG) method are proposed, integrated into the projection method framework in patchwork and overall differentiable manners with MG method, respectively. The solution of the pressure Poisson equation split from the Navier-Stokes equations is first generated either following a random field (e.g. Gaussian random field, GRF, computational complexity O(NlogN), N is the number of spatial points) or physical laws (e.g. a kind of spectra, computational complexity O(NM), M is the number of modes), then the source terms, boundary conditions and initial conditions are constructed via balance of equations, avoiding the difficulties of obtaining high-fidelity training datasets. The feasibility of generated data for training Poisson-NN and WTCNN as well as the acceleration performances of the Poisson-NN embedded method and WTCNN-MG method are validated. The results indicate that even without any DNS data, the generated data can train these two models with excellent generalization and accuracy. The data following physical laws can significantly improve the high-frequency approximation, convergence rate, generalization and accuracy than that generated following GRF. The ML projection method offers significant improvements in computational efficiency. Particularly, the Poisson-NN embedded method achieves an average speed-up of 5.83 times over the traditional MG method, while the WTCNN-MG method offers an even greater average speed-up of 7.03 times, demonstrating impressive acceleration performance.

physics.flu-dyn↗

arXiv2025

Method of data forward generation with partial differential equations for machine learning modeling in fluid mechanics

Artificial intelligence (AI) for fluid mechanics has become attractive topic. High-fidelity data is one of most critical issues for the successful applications of AI in fluid mechanics, however, it is expensively obtained or even inaccessible. This study proposes a high-efficient data forward generation method from the partial differential equations (PDEs). Specifically, the solutions of the PDEs are first generated either following a random field (e.g. Gaussian random field, GRF, computational complexity O(NlogN), N is the number of spatial points) or physical laws (e.g. a kind of spectra, computational complexity O(NM), M is the number of modes), then the source terms, boundary conditions and initial conditions are computed to satisfy PDEs. Thus, the data pairs of source terms, boundary conditions and initial conditions with corresponding solutions of PDEs can be constructed. A Poisson neural network (Poisson-NN) embedded in projection method and a wavelet transform convolutional neuro network (WTCNN) embedded in multigrid numerical simulation for solving incompressible Navier-Stokes equations is respectively proposed. The feasibility of generated data for training Poisson-NN and WTCNN is validated. The results indicate that even without any DNS data, the generated data can train these two models with excellent generalization and accuracy. The data following physical laws can significantly improve the convergence rate, generalization and accuracy than that generated following GRF.

cs.LG↗

arXiv2022

Crop mapping in the small sample/no sample case: an approach using a two-level cascade classifier and integrating domain knowledge

Mapping crops using remote sensing technology is important for food security and land management. Machine learning-based methods has become a popular approach for crop mapping in recent years. However, the key to machine learning, acquiring ample and accurate samples, is usually time-consuming and laborious. To solve this problem, a crop mapping method in the small sample/no sample case that integrating domain knowledge and using a cascaded classification framework that combine a weak classifier learned from samples with strong features and a strong classifier trained by samples with weak feature was proposed. First, based on the domain knowledge of various crops, a low-capacity classifier such as decision tree was applied to acquire those pixels with distinctive features and complete observation sequences as "strong feature" samples. Then, to improve the representativeness of these samples, sample augmentation strategy that artificially remove the observations of "strong feature" samples according to the average valid observation proportion in target area was applied. Finally, based on the original samples and augmented samples, a large-capacity classifier such as random forest was trained for crop mapping. The method achieved an overall accuracy of 82% in the MAP crop recognition competition held by Syngenta Group, China in 2021 (third prize, ranked fourth). This method integrates domain knowledge to overcome the difficulties of sample acquisition, providing a convenient, fast and accurate solution for crop mapping.

cs.CV↗

arXiv2020

Team Performance Evaluation Model based on Network Feature Extraction

Teamwork is increasingly important in today's society. This paper aims at the problem of team performance evaluation. Through complex network feature extraction, we establishes the passing network and team performance evaluation model. Finally, this paper proposes strategy for Huskies team and extend the model to the general team.

cs.SI↗