Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks109 from public data
- RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control4,689
This work proposes a simple, general recipe to enable a single end-to-end trained model to both learn to map robot observations to actions and enjoy the benefits of large-scale pretraining on language and vision-language data from the web.
- Code as Policies: Language Model Programs for Embodied Control2,078
Code as Policies is presented, a robot-centric formulation of language model generated programs (LMPs) that can represent reactive policies (e.g., impedance controllers), as well as waypoint-based policies (vision-based pick and place, trajectory-based control), demonstrated across multiple real robot platforms.
- Multimodal Learning With Transformers: A Survey1,183
This paper presents a comprehensive survey of Transformer techniques oriented at multimodal data, and a summary of the common challenges and designs shared by the multi-modal Transformer models and applications.
- 511
- Structure analysis of soccer video with domain knowledge and hidden Markov models244
This paper defines two mutually exclusive states of the game, play and break based on the rules of soccer, and model the stochastic structures of each state of the game with a set of hidden Markov models.
- RT-Trajectory: Robotic Task Generalization via Hindsight Trajectory Sketches188
This work proposes a policy conditioning method using rough trajectory sketches that is practical, easy to specify, and allows the policy to effectively perform new tasks that would otherwise be challenging to perform, and finds that RT-Trajectory is able to perform a wider range of tasks when provided the same training data.
- RoboVQA: Multimodal Long-Horizon Reasoning for Robotics176
A single video-conditioned model named RoboVQA-VideoCoCa trained on the dataset is demonstrated that is capable of performing a variety of grounded high-level reasoning tasks in broad realistic settings with a cognitive intervention rate 46% lower than the zeroshot state of the art visual language model (VLM) baseline and is able to guide real robots through long-horizon tasks.
- Deep Learning for Free-Hand Sketch: A Survey and A Toolbox160
A comprehensive survey of the deep learning techniques oriented at free-hand sketch data, and the applications that they enable.
- Advances in deep concealed scene understanding148
A comprehensive survey of deep learning techniques aimed at CSU, including a taxonomy, task-specific challenges, and ongoing developments is presented, and the largest and latest benchmark for concealed object segmentation (COS) is offered.
- Fault detection and identification using Bayesian recurrent neural networks147
A novel probabilistic fault detection and identification method is proposed which adopts a newly developed deep learning approach using Bayesian recurrent neural networks~(BRNNs) with variational dropout which yields uncertainty estimates which allow for simultaneous fault detection of chemical processes, direct fault identification, and fault propagation analysis.
- Landscape of the regulatory elements for lysine 2-hydroxyisobutyrylation pathway146
This study shows that Esa1p in budding yeast and its homologue Tip60 in human could add Khib to substrate proteins both in vitro and in vivo, and reports the first global profiling of Khib proteome in mammalian cells, identifying 6 548 Khib sites on 1 725 substrate proteins.
- AutoRT: Embodied Foundation Models for Large Scale Orchestration of Robotic Agents143
This paper proposes AutoRT, a system that leverages existing foundation models to scale up the deployment of operational robots in completely unseen scenarios with minimal human supervision, and leverages vision-language models (VLMs) for scene understanding and grounding, and further uses large language models (LLMs) for proposing diverse and novel instructions to be performed by a fleet of robots.
- SAM struggles in concealed scenes — empirical study on “Segment Anything”134
This report chooses three concealed scenes, i.e., camouflaged animals, industrial defects, and medical lesions, to evaluate SAM under unprompted settings, and observation is that SAM looks unskilled in concealed scenes.
- SketchMate: Deep Hashing for Million-Scale Human Sketch Retrieval130
This work proposes a deep hashing framework for sketch retrieval that, for the first time, works on a multi-million scale human sketch dataset and shows that state-of-the-art hashing models specifically engineered for static images fail to perform well on temporal sketch data.
- Predicting pipeline leakage in petrochemical system through GAN and LSTM108
A novel neural network based classification model using GAN and LSTM using a sliding window and a key feature selecting method in data preparation stage to address the noise and high dimension problems and is more suitable for fault prediction of pipeline leakage in petrochemical system.
- Multigraph Transformer for Free-Hand Sketch Recognition103
This work designs a novel graph neural network (GNN), the multigraph transformer (MGT), for learning representations of sketches from multiple graphs, which simultaneously capture global and local geometric stroke structures as well as temporal information.
- 101
- PortraitBooth: A Versatile Portrait Model for Fast Identity-Preserved Personalization96
Por-traitBooth leverages subject embeddings from aface recognition model for personalized image generation without fine-tuning and eliminates computational overhead and mitigates identity distortion, and introduces dynamic identity preservation strategy, which ensures close resemblance to the original image identity.
- 94
- Cross-modal subspace learning for fine-grained sketch-based image retrieval90
Through thorough examination of the experimental results, it is demonstrated that the subspace learning can effectively model the sketch-photo domain-gap, and a few key insights to drive future research are drawn.
- Decentralized fault detection and diagnosis via sparse PCA based decomposition and Maximum Entropy decision fusion83
Empirical studies on the benchmark Tennessee Eastman Process data demonstrated that the proposed decentralized approach achieves accuracy comparable to that of the fully centralized approach, while offering benefits in terms of fault tolerance, reusability, and scalability.
- Learning Model Predictive Controllers with Real-Time Attention for Real-World Navigation77
A new class of implicit control policies combining the benefits of imitation learning with the robust handling of system constraints from Model Predictive Control (MPC) are presented, called Performer-MPC, which achieves 40% better goal reached in cluttered environments and 65% better on social metrics when navigating around humans.
- Crystal Structure of Enterovirus 71 RNA-Dependent RNA Polymerase Complexed with Its Protein Primer VPg: Implication for a trans Mechanism of VPg Uridylylation74
The crystal structure of EV71 3Dpol complexed with VPg is reported, demonstrating that different picornaviruses use distinct binding sites for VPg uridylylation.
- 69
- BeHere: a VR/SAR remote collaboration system based on virtual replicas sharing gesture and avatar in a procedural task52
A Virtual Reality/Spatial Augmented Reality remote collaboration system based on 3D virtual replicas and sharing gestures and avatar, BeHere, which enables a remote expert in VR to guide a local worker in real-time to complete a procedural task in the real-world.
- Deep Self-Supervised Representation Learning for Free-Hand Sketch52
This paper proposes a set of pretext tasks specifically designed for sketches that mimic different drawing styles, and exploits the use of the textual convolution network together with the convolutional neural network in a dual-branch architecture for sketch feature learning, as means to accommodate the sequential stroke nature of sketches.
- Internet Traffic Classification Using C4.5 Decision Tree52
A new method based on C4.5 decision tree that builds a classification model using information entropy in training data and classifies flows just by a simple search of the decision tree.
- A physiological theory of depth perception from vertical disparity47
A new theory for depth perception from vertical disparity is proposed based on the oriented binocular receptive fields of visual cortical cells and on the radial bias of the preferred-orientation distribution in the cortex that suggests that oriented cells may treat a vertical disparity as a weaker, equivalent horizontal disparity.
- Fine-Grained Instance-Level Sketch-Based Video Retrieval40
This work proposes a novel cross-modal retrieval problem of fine-grained instance-level sketch-based video retrieval (FG-SBVR), where a sketch sequence is used as a query to retrieve a specific target video instance.
- Statistical root cause analysis of novel faults based on digraph models35
A new framework for novel fault diagnosis is proposed, which integrates causal reasoning on signed digraph models with multivariate statistical process monitoring in order to diagnose novel faults whose fault mechanisms and relevant historical data are not available.
- 34
- 33
- Instance-Level Coupled Subspace Learning for Fine-Grained Sketch-Based Image Retrieval27
This paper forms a cross-domain framework specifically designed for the task of FG-SBIR that simultaneously conducts instance-level retrieval and attribute prediction and proposes a joint view selection and attribute subspace learning algorithm to learn domain projection matrices for photo and sketch, respectively.
- LAKE-RED: Camouflaged Images Generation by Latent Background Knowledge Retrieval-Augmented Diffusion25
This paper proposes a Latent Background Knowledge Retrieval-Augmented Diffusion (LAKE-RED) for camouflaged image generation, the first knowledge retrieval-augmented method with interpretability for camouflaged generation, and is the first knowledge retrieval-augmented method with interpretability for camouflaged generation.
- Newton-MR: Newton's Method Without Smoothness or Convexity25
It is shown that two simple modifications of Newton-CG result in an algorithm, called Newton-MR, which offers a diverse range of algorithmic and theoretical advantages and which can be applied, beyond the traditional convex settings, to invex problems.
- 25
- High dynamic range 3D measurement technique based on adaptive fringe projection and curve fitting24
Multiple experimental results show that the algorithm can reconstruct high-quality 3D shapes for highly reflective objects, with high adaptability and reliability in high dynamic range measurements.
- 21
- Frontiers in Intelligent Colonoscopy19
This study investigates the frontiers of intelligent colonoscopy techniques and their prospective implications for multimodal medical applications by assessing the current data-centric and model-centric landscapes through four tasks for colonoscopic scene perception, including classification, detection, segmentation, and vision-language understanding.
- Deep Image Feature Learning With Fuzzy Rules19
The method Deep Image Feature Learning with Fuzzy Rules with Fuzzy Rules (DIFL-FR) is proposed, where human decision process is embedded in feature extraction by combining fuzzy logic rule-based modeling with deep-stacked learning strategy.
- Inexact Nonconvex Newton-Type Methods19
Using certain conditions on such approximations, the paper establishes optimal worst-case iteration complexities as the exact counterparts of nonconvex Newton-type methods.
- Newton-MR: Inexact Newton Method with minimum residual sub-problem solver18
Newton-MR can be readily applied for unconstrained optimization of a class of non-convex problems known as invex, which subsumes convexity as a sub-class, and fast local/global convergence can be guaranteed under a novel inexactness condition, which is much weaker than the prior related works.
- Short-length SSVEP data extension by a novel generative adversarial networks based framework17
The feasibility of the proposed method to extend the data length for short-time SSVEP signals for developing a high-performance BCI system is substantiated and the performance of traditional frequency recognition methods and deep learning-based methods have been significantly improved under limited calibration data.
- 17
- 16
- Sparse Autoregressive Modeling via the Least Absolute LP-Norm Penalized Solution14
A novel ADMM-based AR estimator termed LAPPS (Least Absolute LP (0 < p < 1) Penalized Solution) which employs the L1-loss function for the residual error to alleviate the influence of outliers and another Lp-penalty term to obtain the sparse AR parameters while suppressing any spurious noise that may be present is introduced.
- Improved YOLOv4-tiny based on attention mechanism for skin detection13
The model, which integrates the CBAM attention mechanism and YOLOv4-tiny, achieved a good balance between model size and detection accuracy, and was evaluated using the comprehensive evaluation index.
- Deep Zero-Shot Learning for Scene Sketch13
The augmented semantic vector is proposed to conduct domain alignment by fusing multi-modal semantic knowledge (e.g., cartoon image, natural image, text description), and adopt attention-based network for scene sketch feature learning.
- A continuous learning approach to brain tumor segmentation: integrating multi-scale spatial distillation and pseudo-labeling strategies12
This framework addresses common challenges in brain tumour segmentation, such as computational complexity, limited generalisability, and the extensive need for manual annotation, by exploiting the coordinated capabilities of the ResNet18 and DeepLabV3+ network architectures.
- 12
- 12
- 12
- 9
- MAL-Net: A Multi-Label Deep Learning Framework Integrating LSTM and Multi-Head Attention for Enhanced Classification of IgA Nephropathy Subtypes Using Clinical Sensor Data7
MAL-Net provides a robust solution for multi-label IgAN subtype classification, tackling challenges such as data heterogeneity, class imbalance, and feature interdependencies by integrating clinical sensor data.
- 7
- 7
- 5
- SiCL: Silhouette-Driven Contrastive Learning for Unsupervised Person Re-Identification with Clothes Change4
A semantic mask-driven contrastive learning approach is proposed, in which silhouette masks are embedded into contrastive learning framework as the semantic prompts and cross-clothes invariance is learnt from hierarchically semantic neighbor structure by combining both RGB and semantic features in a two-branches network.
- Robust visual tracking by embedding combination and weighted-gradient optimization4
This paper proposes a negative sample embedding combination network, which helps to learn several sub-embeddings and combine them to build a robust classifier, and proposes a weighted-gradient loss to handle the imbalance between easy and hard samples.
- 4
- 3
- 3
- 3
- 3
- 3
- 3
- 3
- A content based video retrieval method for surveillance and forensic applications3
This paper introduces an ontology-based video retrieval approach, which represents videos with object ontologies and event ontologies, and annotates videos accordingly, and describes a user-friendly interface for querying surveillance videos using event dictionaries.
- 2
- Comparing between virtual reality based pre-clinical implantation training and traditional learning methods2
Compared with traditional PCIT, VR based PCIT has obvious influence on enhancing students’ knowledge mastery, study willingness and learning efficiency.
- 2
- Correction: Optimal Number of Clusters by Measuring Similarity Among Topographies for Spatio-Temporal ERP Analysis2
In “Materials and Methods” section, under the subheading “Simulated ERP Data”, the year 2006 which presented in the third line of the first paragraph should be removed.
- 2
- 2
- 2
- WebCDN: A Peer-to-Peer Web Browser CDN Based WebRTC2
The design and implementation of the WebCDN system is described in terms of both server and client side in detail, which achieves a content distribution network by web users and greatly reduces the network traffic with acceptable service latency.
- PegasusFlow: Parallel Rolling-Denoising Score Sampling for Robot Diffusion Planner Flow Matching1
PegasusFlow is introduced, a parallel rolling-denoising framework that enables direct sampling of trajectory score gradients from environmental interaction, completely bypassing the need for expert data and significantly outperforms baselines.
- 1
- 1
- 1
- 1
- 1
- 1
- Coordination in Agile Software Projects1
This research investigates the coordination mechanisms used in agile software development and the factors that impact their effectiveness, and distinguishes coordination mechanisms as vertical coordination and horizontal coordination.
- –
- –
- –
- –
- A Study on the Postgraduate Quality Evaluation Model and Its Recognition Analysis–
Investigating the degree of acceptance among diverse groups—such as current undergraduate students, current postgraduate students, graduates with less than three years of work experience, and those with more than three years of work experience—can offer valuable insights for improving and optimizing postgraduate education and training practices.
- –
- –
- –
- Text Mining Based on Web Crawler Technology to Analyze the “Reacting Force” of Film Reviews on Films–
- –
- –
- Design and Implementation of Academic Search Aggregation Engine–
This thesis designs and implements an academic search aggregation engine based on cloud computing and some clustering algorithms and analyzes the needs of the researcher and other customers for the paper search, and proposes a system design that distribute the search request to multiple academic search engines and then aggregated results.
- –
- –
- –
- Analysis and improvement of secueity services schemes in EPON system–
From the analysis of various types of potential network structure targeted attacks in Ethernet passive optical network, EPON, the results demonstrate the efficiency of the indicated security schemes which also meet the security condition of non-renouncement.
- –
- –
- Indemnificatory research on key services of survivable networks–
This paper proposes MESMMA (A Model for Essential Services Mobility Based on Mobile Agent), a Mobile Agent mobility-based model used to ensure the nonstop running of key services of survivable networks.
- Introduction and Explanation to IMS E-Portfolio Specification–
The author introduces the concepts of usability and usability testing, the process of usability testing and principles of usabilityTesting on LMS software with some example on Sakai LMS.
- The Method of Improving Sports Information Retrieval Efficiency on Network–
Sports network for information resources complex disorder, lack of depth, the relative dispersion characteristics such as search words to choose from, professional site selection and the use of advanced search techniques to increase the efficiency of the search.
- –
- A method of γ spectrum identification based on the BP neural network–
The technique of identifying gamma spectrums using BP neural network is discussed and factors influencing the identification results are approached in this paper.
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-14. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.