arXiv subjects

Yinghao Cai

Publications and source records attributed to Yinghao Cai.

18 recordsLinked to original sources

arXiv2026

FeelWorld: Visuo-Tactile World Model for Hierarchical Contact Prediction and Planning

Humans plan physical interactions by imagining the possible outcomes of candidate actions. However, existing visual world models primarily capture appearance dynamics while overlooking the tactile states that govern contact-rich interactions, potentially producing imagined futures that appear visually plausible but violate physical dynamics. We introduce FeelWorld, a hierarchical visuo-tactile world model that jointly predicts future visual latents and three tactile states. FeelWorld organizes these states hierarchically as contact state, a 3D tactile latent that encodes force-related information, and slip state. These states are jointly predicted by a shared latent dynamics model with explicit supervision. To prevent irrelevant tactile signals during free-space motion from degrading visual prediction, we introduce a contact-gated asymmetric attention mechanism that maintains a visual-only prediction pathway before contact and enables joint visuo-tactile dynamics prediction during contact. The model is further trained with autoregressive rollouts and context noise injection to improve robustness to compounding errors. The predicted contact and slip states also support contact-aware CEM planning. Experiments on chip grasping, fruit grasping, and USB insertion show that FeelWorld reduces 10-step LPIPS from 0.084 to 0.058 and maintains an LPIPS that is 61% lower than that of the visual baseline after an 80-step autoregressive rollout. FeelWorld also achieves an average zero-shot planning success rate of 81.7%, providing an effective approach for incorporating tactile sensing into world models.

cs.RO↗

arXiv2026

CoRe: Combined Rewards with Vision-Language Model Feedback for Preference-Aligned Reinforcement Learning

Reward design remains a central challenge in reinforcement learning (RL). Hand-crafted rewards are often difficult to specify and may lead to suboptimal policies, while learned rewards from preferences can suffer from inefficiency and unstable training. Inspired by the dual nature of human learning explored in cognitive science, we decompose rewards into two complementary components: Formal Rewards (FR), explicitly designed based on task knowledge, and Residual Rewards (RR), learned from observations to capture implicit and nuanced preferences. Based on this decomposition, we propose CoRe, a hybrid framework that integrates FR and RR with vision-language models (VLMs) feedback to achieve preference-aligned policies without human involvement. Our contributions are twofold: (1) We propose a Formal Reward Module (FRM) that leverages VLMs to iteratively design and optimize FR based on task knowledge and preference feedback, enabling the continual improvement of policy during training; (2) We introduce a Residual Reward Module (RRM) that learns RR from video-level preference by employing VLMs to generate preference labels and capturing nuanced rewards that complement FR, ensuring alignment with human intent. Through the synergy of FRM and RRM, CoRe enables the automatic construction of reliable rewards that are efficient and preference-aligned. Extensive experiments demonstrate that CoRe outperforms existing approaches in terms of policy learning effectiveness and efficiency on ten robotic manipulation tasks in simulation and five real-worlds. Videos can be found on our project website: https://core-2026.github.io/

cs.RO↗

arXiv2026

FG-CLTP: Fine-Grained Contrastive Language Tactile Pretraining for Robotic Manipulation

Recent advancements in integrating tactile sensing into vision-language-action (VLA) models have demonstrated transformative potential for robotic perception. However, existing tactile representations predominantly rely on qualitative descriptors (e.g., texture), neglecting quantitative contact states such as force magnitude, contact geometry, and principal axis orientation, which are indispensable for fine-grained manipulation. To bridge this gap, we propose FG-CLTP, a fine-grained contrastive language tactile pretraining framework. We first introduce a novel dataset comprising over 100k tactile 3D point cloud-language pairs that explicitly capture multidimensional contact states from the sensor's perspective. We then implement a discretized numerical tokenization mechanism to achieve quantitative-semantic alignment, effectively injecting explicit physical metrics into the multimodal feature space. The proposed FG-CLTP model yields a 95.9% classification accuracy and reduces the regression error (MAE) by 52.6% compared to state-of-the-art methods. Furthermore, the integration of 3D point cloud representations establishes a sensor-agnostic foundation with a minimal sim-to-real gap of 3.5%. Building upon this fine-grained representation, we develop a 3D tactile-language-action (3D-TLA) architecture driven by a flow matching policy to enable multimodal reasoning and control. Extensive experiments demonstrate that our framework significantly outperforms strong baselines in contact-rich manipulation tasks, providing a robust and generalizable foundation for tactile-language-action models.

cs.RO↗

arXiv2026

PEAfowl: Perception-Enhanced Multi-View Vision-Language-Action for Bimanual Manipulation

Bimanual manipulation in cluttered scenes requires policies that remain stable under occlusions, viewpoint changes and scene variations. Existing vision-language-action models often lack such robustness because (i) multi-view features are fused via view-agnostic token concatenation, yielding limited cross-view spatial representations, and (ii) language is injected as global conditioning, resulting in coarse instruction grounding. In this paper, we introduce PEAfowl, a perception-enhanced multi-view VLA policy for bimanual manipulation. For spatial perception, PEAfowl predicts per-token depth distributions, performs differentiable 3D lifting, and aggregates local cross-view neighbors to form geometrically grounded, cross-view aligned representations. For language utilization, we propose to replace global conditioning with a Perceiver-style text-aware readout over frozen CLIP visual features, enabling iterative evidence accumulation. To better exploit commodity RGB-D sensing despite noisy and incomplete depth, PEAfowl's depth-distribution lifting naturally supports training-only depth distillation, where a pretrained depth teacher supervises the depth-distribution head to inject refined geometric priors without adding inference overhead. On RoboTwin 2.0 under domain-randomized setting, PEAfowl improves the strongest baseline by 23.0 pp in success rate, and physical experiments further demonstrate improved performance on the evaluated real-robot tasks. Project website: https://peafowlvla.github.io/.

cs.CV↗

arXiv2025

Seeing Through the Brain: New Insights from Decoding Visual Stimuli with fMRI

Understanding how the brain encodes visual information is a central challenge in neuroscience and machine learning. A promising approach is to reconstruct visual stimuli, essentially images, from functional Magnetic Resonance Imaging (fMRI) signals. This involves two stages: transforming fMRI signals into a latent space and then using a pretrained generative model to reconstruct images. The reconstruction quality depends on how similar the latent space is to the structure of neural activity and how well the generative model produces images from that space. Yet, it remains unclear which type of latent space best supports this transformation and how it should be organized to represent visual stimuli effectively. We present two key findings. First, fMRI signals are more similar to the text space of a language model than to either a vision based space or a joint text image space. Second, text representations and the generative model should be adapted to capture the compositional nature of visual stimuli, including objects, their detailed attributes, and relationships. Building on these insights, we propose PRISM, a model that Projects fMRI sIgnals into a Structured text space as an interMediate representation for visual stimuli reconstruction. It includes an object centric diffusion module that generates images by composing individual objects to reduce object detection errors, and an attribute relationship search module that automatically identifies key attributes and relationships that best align with the neural activity. Extensive experiments on real world datasets demonstrate that our framework outperforms existing methods, achieving up to an 8% reduction in perceptual loss. These results highlight the importance of using structured text as the intermediate space to bridge fMRI signals and image reconstruction.

cs.CV↗

arXiv2025

MISCGrasp: Leveraging Multiple Integrated Scales and Contrastive Learning for Enhanced Volumetric Grasping

Robotic grasping faces challenges in adapting to objects with varying shapes and sizes. In this paper, we introduce MISCGrasp, a volumetric grasping method that integrates multi-scale feature extraction with contrastive feature enhancement for self-adaptive grasping. We propose a query-based interaction between high-level and low-level features through the Insight Transformer, while the Empower Transformer selectively attends to the highest-level features, which synergistically strikes a balance between focusing on fine geometric details and overall geometric structures. Furthermore, MISCGrasp utilizes multi-scale contrastive learning to exploit similarities among positive grasp samples, ensuring consistency across multi-scale features. Extensive experiments in both simulated and real-world environments demonstrate that MISCGrasp outperforms baseline and variant methods in tabletop decluttering tasks. More details are available at https://miscgrasp.github.io/.

cs.RO↗

arXiv2025

SENIOR: Efficient Query Selection and Preference-Guided Exploration in Preference-based Reinforcement Learning

Preference-based Reinforcement Learning (PbRL) methods provide a solution to avoid reward engineering by learning reward models based on human preferences. However, poor feedback- and sample- efficiency still remain the problems that hinder the application of PbRL. In this paper, we present a novel efficient query selection and preference-guided exploration method, called SENIOR, which could select the meaningful and easy-to-comparison behavior segment pairs to improve human feedback-efficiency and accelerate policy learning with the designed preference-guided intrinsic rewards. Our key idea is twofold: (1) We designed a Motion-Distinction-based Selection scheme (MDS). It selects segment pairs with apparent motion and different directions through kernel density estimation of states, which is more task-related and easy for human preference labeling; (2) We proposed a novel preference-guided exploration method (PGE). It encourages the exploration towards the states with high preference and low visits and continuously guides the agent achieving the valuable samples. The synergy between the two mechanisms could significantly accelerate the progress of reward and policy learning. Our experiments show that SENIOR outperforms other five existing methods in both human feedback-efficiency and policy convergence speed on six complex robot manipulation tasks from simulation and four real-worlds. Videos can be found on our project website: https://2025senior.github.io/

cs.RO↗

arXiv2025

CLTP: Contrastive Language-Tactile Pre-training for 3D Contact Geometry Understanding

Recent advancements in integrating tactile sensing with vision-language models (VLMs) have demonstrated remarkable potential for robotic multimodal perception. However, existing tactile descriptions remain limited to superficial attributes like texture, neglecting critical contact states essential for robotic manipulation. To bridge this gap, we propose CLTP, an intuitive and effective language tactile pretraining framework that aligns tactile 3D point clouds with natural language in various contact scenarios, thus enabling contact-state-aware tactile language understanding for contact-rich manipulation tasks. We first collect a novel dataset of 50k+ tactile 3D point cloud-language pairs, where descriptions explicitly capture multidimensional contact states (e.g., contact location, shape, and force) from the tactile sensor's perspective. CLTP leverages a pre-aligned and frozen vision-language feature space to bridge holistic textual and tactile modalities. Experiments validate its superiority in three downstream tasks: zero-shot 3D classification, contact state classification, and tactile 3D large language model (LLM) interaction. To the best of our knowledge, this is the first study to align tactile and language representations from the contact state perspective for manipulation tasks, providing great potential for tactile-language-action model learning. Code and datasets are open-sourced at https://sites.google.com/view/cltp/.

cs.RO↗

arXiv2025

NeuGrasp: Generalizable Neural Surface Reconstruction with Background Priors for Material-Agnostic Object Grasp Detection

Robotic grasping in scenes with transparent and specular objects presents great challenges for methods relying on accurate depth information. In this paper, we introduce NeuGrasp, a neural surface reconstruction method that leverages background priors for material-agnostic grasp detection. NeuGrasp integrates transformers and global prior volumes to aggregate multi-view features with spatial encoding, enabling robust surface reconstruction in narrow and sparse viewing conditions. By focusing on foreground objects through residual feature enhancement and refining spatial perception with an occupancy-prior volume, NeuGrasp excels in handling objects with transparent and specular surfaces. Extensive experiments in both simulated and real-world scenarios show that NeuGrasp outperforms state-of-the-art methods in grasping while maintaining comparable reconstruction quality. More details are available at https://neugrasp.github.io/.

cs.RO↗

arXiv2024

Exploring Consistency in Graph Representations:from Graph Kernels to Graph Neural Networks

Graph Neural Networks (GNNs) have emerged as a dominant approach in graph representation learning, yet they often struggle to capture consistent similarity relationships among graphs. While graph kernel methods such as the Weisfeiler-Lehman subtree (WL-subtree) and Weisfeiler-Lehman optimal assignment (WLOA) kernels are effective in capturing similarity relationships, they rely heavily on predefined kernels and lack sufficient non-linearity for more complex data patterns. Our work aims to bridge the gap between neural network methods and kernel approaches by enabling GNNs to consistently capture relational structures in their learned representations. Given the analogy between the message-passing process of GNNs and WL algorithms, we thoroughly compare and analyze the properties of WL-subtree and WLOA kernels. We find that the similarities captured by WLOA at different iterations are asymptotically consistent, ensuring that similar graphs remain similar in subsequent iterations, thereby leading to superior performance over the WL-subtree kernel. Inspired by these findings, we conjecture that the consistency in the similarities of graph representations across GNN layers is crucial in capturing relational structures and enhancing graph classification performance. Thus, we propose a loss to enforce the similarity of graph representations to be consistent across different layers. Our empirical analysis verifies our conjecture and shows that our proposed consistency loss can significantly enhance graph classification performance across several GNN backbones on various datasets.

cs.LG↗

arXiv2023

Sparse and Low-Rank High-Order Tensor Regression via Parallel Proximal Method

Recently, tensor data (or multidimensional array) have been generated in many modern applications, such as functional magnetic resonance imaging (fMRI) in neuroscience and videos in video analysis. Many efforts are made in recent years to predict the relationship between tensor features and univariate responses. However, previously proposed methods either lose structural information within tensor data or have prohibitively expensive time costs, especially for large-scale data with high-order structures. To address such problems, we propose the Sparse and Low-rank Tensor Regression (SLTR) model. Our model enforces sparsity and low-rankness of the tensor coefficient by directly applying $\ell_1$ norm and tensor nuclear norm, such that it preserves structural information of the tensor. To make the solving procedure scalable and efficient, SLTR makes use of the proximal gradient method, which can be easily implemented parallelly. We evaluate SLTR on several simulated datasets and one video action recognition dataset. Experiment results show that, compared with previous models, SLTR can obtain a better solution with much fewer time costs. Moreover, our model's predictions exhibit meaningful interpretations on the video dataset.

cs.LG↗

arXiv2022

Real-Time Marker Localization Learning for GelStereo Tactile Sensing

Visuotactile sensing technology is becoming more popular in tactile sensing, but the effectiveness of the existing marker detection localization methods remains to be further explored. Instead of contour-based blob detection, this paper presents a learning-based marker localization network for GelStereo visuotactile sensing called Marknet. Specifically, the Marknet presents a grid regression architecture to incorporate the distribution of the GelStereo markers. Furthermore, a marker rationality evaluator (MRE) is modelled to screen suitable prediction results. The experimental results show that the Marknet combined with MRE achieves 93.90% precision for irregular markers in contact areas, which outperforms the traditional contour-based blob detection method by a large margin of 42.32%. Meanwhile, the proposed learning-based marker localization method can achieve better real-time performance beyond the blob detection interface provided by the OpenCV library through GPU acceleration, which we believe will lead to considerable perceptual sensitivity gains in various robotic manipulation tasks.

cs.RO↗

arXiv2022

Learning Symbolic Operators: A Neurosymbolic Solution for Autonomous Disassembly of Electric Vehicle Battery

The booming of electric vehicles demands efficient battery disassembly for recycling to be environment-friendly. Currently, battery disassembly is still primarily done by humans, probably assisted by robots, due to the unstructured environment and high uncertainties. It is highly desirable to design autonomous solutions to improve work efficiency and lower human risks in high voltage and toxic environments. This paper proposes a novel neurosymbolic method, which augments the traditional Variational Autoencoder (VAE) model to learn symbolic operators based on raw sensory inputs and their relationships. The symbolic operators include a probabilistic state symbol grounding model and a state transition matrix for predicting states after each execution to enable autonomous task and motion planning. At last, the method's feasibility is verified through test results.

cs.RO↗

arXiv2020

ACDER: Augmented Curiosity-Driven Experience Replay

Exploration in environments with sparse feedback remains a challenging research problem in reinforcement learning (RL). When the RL agent explores the environment randomly, it results in low exploration efficiency, especially in robotic manipulation tasks with high dimensional continuous state and action space. In this paper, we propose a novel method, called Augmented Curiosity-Driven Experience Replay (ACDER), which leverages (i) a new goal-oriented curiosity-driven exploration to encourage the agent to pursue novel and task-relevant states more purposefully and (ii) the dynamic initial states selection as an automatic exploratory curriculum to further improve the sample-efficiency. Our approach complements Hindsight Experience Replay (HER) by introducing a new way to pursue valuable states. Experiments conducted on four challenging robotic manipulation tasks with binary rewards, including Reach, Push, Pick&Place and Multi-step Push. The empirical results show that our proposed method significantly outperforms existing methods in the first three basic tasks and also achieves satisfactory performance in multi-step robotic task learning.

cs.RO↗

arXiv2019

Tutorial: Complexity analysis of Singular Value Decomposition and its variants

We compared the regular Singular Value Decomposition (SVD), truncated SVD, Krylov method and Randomized PCA, in terms of time and space complexity. It is well-known that Krylov method and Randomized PCA only performs well when k << n, i.e. the number of eigenpair needed is far less than that of matrix size. We compared them for calculating all the eigenpairs. We also discussed the relationship between Principal Component Analysis and SVD.

math.NA↗

arXiv2019

Localizing Discriminative Visual Landmarks for Place Recognition

We address the problem of visual place recognition with perceptual changes. The fundamental problem of visual place recognition is generating robust image representations which are not only insensitive to environmental changes but also distinguishable to different places. Taking advantage of the feature extraction ability of Convolutional Neural Networks (CNNs), we further investigate how to localize discriminative visual landmarks that positively contribute to the similarity measurement, such as buildings and vegetations. In particular, a Landmark Localization Network (LLN) is designed to indicate which regions of an image are used for discrimination. Detailed experiments are conducted on open source datasets with varied appearance and viewpoint changes. The proposed approach achieves superior performance against state-of-the-art methods.

cs.CV↗

arXiv2019

Hindsight Generative Adversarial Imitation Learning

Compared to reinforcement learning, imitation learning (IL) is a powerful paradigm for training agents to learn control policies efficiently from expert demonstrations. However, in most cases, obtaining demonstration data is costly and laborious, which poses a significant challenge in some scenarios. A promising alternative is to train agent learning skills via imitation learning without expert demonstrations, which, to some extent, would extremely expand imitation learning areas. To achieve such expectation, in this paper, we propose Hindsight Generative Adversarial Imitation Learning (HGAIL) algorithm, with the aim of achieving imitation learning satisfying no need of demonstrations. Combining hindsight idea with the generative adversarial imitation learning (GAIL) framework, we realize implementing imitation learning successfully in cases of expert demonstration data are not available. Experiments show that the proposed method can train policies showing comparable performance to current imitation learning methods. Further more, HGAIL essentially endows curriculum learning mechanism which is critical for learning policies.

cs.LG↗

arXiv2015

The Multi-Strand Graph for a PTZ Tracker

High-resolution images can be used to resolve matching ambiguities between trajectory fragments (tracklets), which is one of the main challenges in multiple target tracking. A PTZ camera, which can pan, tilt and zoom, is a powerful and efficient tool that offers both close-up views and wide area coverage on demand. The wide-area view makes it possible to track many targets while the close-up view allows individuals to be identified from high-resolution images of their faces. A central component of a PTZ tracking system is a scheduling algorithm that determines which target to zoom in on. In this paper we study this scheduling problem from a theoretical perspective, where the high resolution images are also used for tracklet matching. We propose a novel data structure, the Multi-Strand Tracking Graph (MSG), which represents the set of tracklets computed by a tracker and the possible associations between them. The MSG allows efficient scheduling as well as resolving -- directly or by elimination -- matching ambiguities between tracklets. The main feature of the MSG is the auxiliary data saved in each vertex, which allows efficient computation while avoiding time-consuming graph traversal. Synthetic data simulations are used to evaluate our scheduling algorithm and to demonstrate its superiority over a naïve one.

cs.CV↗