This article describes a multi-modal method using simulated Lidar data via ray tracing and image pixel loss with differentiable rendering to optimize an object's position with respect to an observer or some referential objects in a computer graphics scene. Object position optimization is completed using gradient descent with the loss function being influenced by both modalities. Typical object placement optimization is done using image pixel loss with differentiable rendering only, this work shows the use of a second modality (Lidar) leads to faster convergence. This method of fusing sensor input presents a potential usefulness for autonomous vehicles, as these methods can be used to establish the locations of multiple actors in a scene. This article also presents a method for the simulation of multiple types of data to be used in the training of autonomous vehicles.
Environmental, Social, and Governance (ESG) has been used as a metric to measure the negative impacts and enhance positive outcomes of companies in areas such as the environment, society, and governance. Recently, investors have increasingly recognized the significance of ESG criteria in their investment choices, leading businesses to integrate ESG principles into their operations and strategies. The Multi-Lingual ESG Issue Identification (ML-ESG) shared task encompasses the classification of news documents into 35 distinct ESG issue labels. In this study, we explored multiple strategies harnessing BERT language models to achieve accurate classification of news documents across these labels. Our analysis revealed that the RoBERTa classifier emerged as one of the most successful approaches, securing the second-place position for the English test dataset, and sharing the fifth-place position for the French test dataset. Furthermore, our SVM-based binary model tailored for the Chinese language exhibited exceptional performance, earning the second-place rank on the test dataset.
The Transformer architecture has two main non-embedding components: Attention and the Feed Forward Network (FFN). Attention captures interdependencies between words regardless of their position, while the FFN non-linearly transforms each input token independently. In this work we explore the role of the FFN, and find that despite taking up a significant fraction of the model's parameters, it is highly redundant. Concretely, we are able to substantially reduce the number of parameters with only a modest drop in accuracy by removing the FFN on the decoder layers and sharing a single FFN across the encoder. Finally we scale this architecture back to its original size by increasing the hidden dimension of the shared FFN, achieving substantial gains in both accuracy and latency with respect to the original Transformer Big.
Recently, transformer-based methods have dominated 3D instance segmentation, where mask attention is commonly involved. Specifically, object queries are guided by the initial instance masks in the first cross-attention, and then iteratively refine themselves in a similar manner. However, we observe that the mask-attention pipeline usually leads to slow convergence due to low-recall initial instance masks. Therefore, we abandon the mask attention design and resort to an auxiliary center regression task instead. Through center regression, we effectively overcome the low-recall issue and perform cross-attention by imposing positional prior. To reach this goal, we develop a series of position-aware designs. First, we learn a spatial distribution of 3D locations as the initial position queries. They spread over the 3D space densely, and thus can easily capture the objects in a scene with a high recall. Moreover, we present relative position encoding for the cross-attention and iterative refinement for more accurate position queries. Experiments show that our approach converges 4x faster than existing work, sets a new state of the art on ScanNetv2 3D instance segmentation benchmark, and also demonstrates superior performance across various datasets. Code and models are available at https://github.com/dvlab-research/Mask-Attention-Free-Transformer.
This paper presents an approach to automatically annotate automotive radar data with AI-segmented aerial camera images. For this, the images of an unmanned aerial vehicle (UAV) above a radar vehicle are panoptically segmented and mapped in the ground plane onto the radar images. The detected instances and segments in the camera image can then be applied directly as labels for the radar data. Owing to the advantageous bird's eye position, the UAV camera does not suffer from optical occlusion and is capable of creating annotations within the complete field of view of the radar. The effectiveness and scalability are demonstrated in measurements, where 589 pedestrians in the radar data were automatically labeled within 2 minutes.
Rotary Position Embeddings (RoPE) have been shown to effectively encode positional information in transformer-based language models. However, these models fail to generalize past the sequence length they were trained on. We present YaRN (Yet another RoPE extensioN method), a compute-efficient method to extend the context window of such models, requiring 10x less tokens and 2.5x less training steps than previous methods. Using YaRN, we show that LLaMA models can effectively utilize and extrapolate to context lengths much longer than their original pre-training would allow, while also surpassing previous the state-of-the-art at context window extension. In addition, we demonstrate that YaRN exhibits the capability to extrapolate beyond the limited context of a fine-tuning dataset. The models fine-tuned using YaRN has been made available and reproduced online up to 128k context length at https://github.com/jquesnelle/yarn
This paper presents a neural-enhanced probabilistic model and corresponding factor graph-based sum-product algorithm for robust localization and tracking in multipath-prone environments. The introduced hybrid probabilistic model consists of physics-based and data-driven measurement models capturing the information contained in both, the line-of-sight (LOS) component as well as in multipath components (NLOS components). The physics-based and data-driven models are embedded in a joint Bayesian framework allowing to derive from first principles a factor graph-based algorithm that fuses the information of these models. The proposed algorithm uses radio signal measurements from multiple base stations to robustly estimate the mobile agent's position together with all model parameters. It provides high localization accuracy by exploiting the position-related information of the LOS component via the physics-based model and robustness by exploiting the geometric imprint of multipath components independent of the propagation channel via the data-driven model. In a challenging numerical experiment involving obstructed LOS situations to all anchors, we show that the proposed sequential algorithm significantly outperforms state-of-the-art methods and attains the posterior Cramer-Rao lower bound even with training data limited to local regions.
After an earthquake, it is particularly important to provide the necessary resources on site because a large number of infrastructures need to be repaired or newly constructed. Due to the complex construction environment after the disaster, there are potential safety hazards for human labors working in this environment. With the advancement of robotic technology and artificial intelligent (AI) algorithms, smart robotic technology is the potential solution to provide construction resources after an earthquake. In this paper, the robotic crane with advanced AI algorithms is proposed to provide resources for infrastructure reconstruction after an earthquake. The proximal policy optimization (PPO), a reinforcement learning (RL) algorithm, is implemented for 3D lift path planning when transporting the construction materials. The state and reward function are designed in detail for RL model training. Two models are trained through a loading task in different environments by using PPO algorithm, one considering the influence of obstacles and the other not considering obstacles. Then, the two trained models are compared and evaluated through an unloading task and a loading task in simulation environments. For each task, two different cases are considered. One is that there is no obstacle between the initial position where the construction material is lifted and the target position, and the other is that there are obstacles between the initial position and the target position. The results show that the model that considering the obstacles during training can generate proper actions for the robotic crane to execute so that the crane can automatically transport the construction materials to the desired location with swing suppression, short time consumption and collision avoidance.
Due to the nonlinearities and operational constraints typical to quadcopter missions, Model Predictive Control (MPC) encounters the major challenge of high computational power necessary for the online implementation. This problem may prove impractical, especially for a hardware-limited or small-scale setup. By removing the need for online solvers while keeping the constraint satisfaction and optimality, Explicit MPC (ExMPC) stands out as a strong candidate for this application. Yet, the formulation was usually hindered by the two main problems: nonlinearity and dimensionality. In this paper, we propose an ExMPC solution for the quadcopter position stabilization by analyzing its description (dynamics and constraints) in the flat output space. With the former issue, the system is exactly linearized into a concatenation of three double integrators at a price of cumbersome constraints in the new coordinates. For the latter, with a suitable characterization of these constraints, the stabilizing ExMPC can be computed for each double integrator separately. The controller is then validated via simulations and experimental tests. The proposed scheme achieves similar performance and guarantees to the state-of-the-art solution but with notably less computational effort, allowing scalability in a centralized manner
Text-based Visual Question Answering (TextVQA) aims at answering questions about the text in images. Most works in this field focus on designing network structures or pre-training tasks. All these methods list the OCR texts in reading order (from left to right and top to bottom) to form a sequence, which is treated as a natural language ``sentence''. However, they ignore the fact that most OCR words in the TextVQA task do not have a semantical contextual relationship. In addition, these approaches use 1-D position embedding to construct the spatial relation between OCR tokens sequentially, which is not reasonable. The 1-D position embedding can only represent the left-right sequence relationship between words in a sentence, but not the complex spatial position relationship. To tackle these problems, we propose a novel method named Separate and Locate (SaL) that explores text contextual cues and designs spatial position embedding to construct spatial relations between OCR texts. Specifically, we propose a Text Semantic Separate (TSS) module that helps the model recognize whether words have semantic contextual relations. Then, we introduce a Spatial Circle Position (SCP) module that helps the model better construct and reason the spatial position relationships between OCR texts. Our SaL model outperforms the baseline model by 4.44% and 3.96% accuracy on TextVQA and ST-VQA datasets. Compared with the pre-training state-of-the-art method pre-trained on 64 million pre-training samples, our method, without any pre-training tasks, still achieves 2.68% and 2.52% accuracy improvement on TextVQA and ST-VQA. Our code and models will be released at https://github.com/fangbufang/SaL.
The incorporation of 3D data in facial analysis tasks has gained popularity in recent years. Though it provides a more accurate and detailed representation of the human face, accruing 3D face data is more complex and expensive than 2D face images. Either one has to rely on expensive 3D scanners or depth sensors which are prone to noise. An alternative option is the reconstruction of 3D faces from uncalibrated 2D images in an unsupervised way without any ground truth 3D data. However, such approaches are computationally expensive and the learned model size is not suitable for mobile or other edge device applications. Predicting dense 3D landmarks over the whole face can overcome this issue. As there is no public dataset available containing dense landmarks, we propose a pipeline to create a dense keypoint training dataset containing 520 key points across the whole face from an existing facial position map data. We train a lightweight MobileNet-based regressor model with the generated data. As we do not have access to any evaluation dataset with dense landmarks in it we evaluate our model against the 68 keypoint detection task. Experimental results show that our trained model outperforms many of the existing methods in spite of its lower model size and minimal computational cost. Also, the qualitative evaluation shows the efficiency of our trained models in extreme head pose angles as well as other facial variations and occlusions.
Recently, the use of multi-rotor (MR) unmanned aerial vehicles (UAVs) has emerged as a promising solution for establishing flexible free-space optical communication links. We address, in this paper, the accurate channel modeling to assess the benefits of MR UAV-based deployment for such links. In particular, in the absence of active tracking sub- systems, we derive statistical models for ground-to-UAV, UAV- to-UAV, and UAV-to-ground links over both Gamma–Gamma and log-normal atmospheric turbulence models. Unlike previous works on this topic, our proposed model considers the joint effect of atmospheric turbulence along with position and angle-of- arrival fluctuations. The high accuracy of the proposed analytical models is verified by comparing numerically solved and Monte Carlo simulation results in terms of link outage probability. We further discuss the impact of different transmitter/receiver parameters an
In this paper, we propose first a mmWave channel tracking algorithm based on multidimensional orthogonal matching pursuit algorithm (MOMP) using reduced sparsifying dictionaries, which exploits information from channel estimates in previous frames. Then, we present an algorithm to obtain the vehicle's initial location for the current frame by solving a system of geometric equations that leverage the estimated path parameters. Next, we design an attention network that analyzes the series of channel estimates, the vehicle's trajectory, and the initial estimate of the position associated with the current frame, to generate a refined, high accuracy position estimate. The proposed system is evaluated through numerical experiments using realistic mmWave channel series generated by ray-tracing. The experimental results show that our system provides a 2D position tracking error below 20 cm, significantly outperforming previous work based on Bayesian filtering.
Vehicular communication networks are rapidly emerging as vehicles become smarter. However, these networks are increasingly susceptible to various attacks. The situation is exacerbated by the rise in automated vehicles complicates, emphasizing the need for security and authentication measures to ensure safe and effective traffic management. In this paper, we propose a novel hybrid physical layer security (PLS)-machine learning (ML) authentication scheme by exploiting the position of the transmitter vehicle as a device fingerprint. We use a time-of-arrival (ToA) based localization mechanism where the ToA is estimated at roadside units (RSUs), and the coordinates of the transmitter vehicle are extracted at the base station (BS).Furthermore, to track the mobility of the moving legitimate vehicle, we use ML model trained on several system parameters. We try two ML models for this purpose, i.e., support vector regression and decision tree. To evaluate our scheme, we conduct binary hypothesis testing on the estimated positions with the help of the ground truths provided by the ML model, which classifies the transmitter node as legitimate or malicious. Moreover, we consider the probability of false alarm and the probability of missed detection as performance metrics resulting from the binary hypothesis testing, and mean absolute error (MAE), mean square error (MSE), and coefficient of determination $\text{R}^2$ to further evaluate the ML models. We also compare our scheme with a baseline scheme that exploits the angle of arrival at RSUs for authentication. We observe that our proposed position-based mechanism outperforms the baseline scheme significantly in terms of missed detections.
In this paper, we introduce Handwritten augmentation, a new data augmentation for handwritten character images. This method focuses on augmenting handwritten image data by altering the shape of input characters in training. The proposed handwritten augmentation is similar to position augmentation, color augmentation for images but a deeper focus on handwritten characters. Handwritten augmentation is data-driven, easy to implement, and can be integrated with CNN-based optical character recognition models. Handwritten augmentation can be implemented along with commonly used data augmentation techniques such as cropping, rotating, and yields better performance of models for handwritten image datasets developed using optical character recognition methods.
The transition to agile manufacturing, Industry 4.0, and high-mix-low-volume tasks require robot programming solutions that are flexible. However, most deployed robot solutions are still statically programmed and use stiff position control, which limit their usefulness. In this paper, we show how a single robot skill that utilizes knowledge representation, task planning, and automatic selection of skill implementations based on the input parameters can be executed in different contexts. We demonstrate how the skill-based control platform enables this with contact-rich wiping tasks on different robot systems. To achieve that in this case study, our approach needs to address different kinematics, gripper types, vendors, and fundamentally different control interfaces. We conducted the experiments with a mobile platform that has a Universal Robots UR5e 6 degree-of-freedom robot arm with position control and a 7 degree-of-freedom KUKA iiwa with torque control.
Recently, Multimodal Large Language Models (MLLMs) that enable Large Language Models (LLMs) to interpret images through visual instruction tuning have achieved significant success. However, existing visual instruction tuning methods only utilize image-language instruction data to align the language and image modalities, lacking a more fine-grained cross-modal alignment. In this paper, we propose Position-enhanced Visual Instruction Tuning (PVIT), which extends the functionality of MLLMs by integrating an additional region-level vision encoder. This integration promotes a more detailed comprehension of images for the MLLM. In addition, to efficiently achieve a fine-grained alignment between the vision modules and the LLM, we design multiple data generation strategies to construct an image-region-language instruction dataset. Finally, we present both quantitative experiments and qualitative analysis that demonstrate the superiority of the proposed model. Code and data will be released at https://github.com/PVIT-official/PVIT.
This paper presents a novel neural implicit radiance representation for free viewpoint relighting from a small set of unstructured photographs of an object lit by a moving point light source different from the view position. We express the shape as a signed distance function modeled by a multi layer perceptron. In contrast to prior relightable implicit neural representations, we do not disentangle the different reflectance components, but model both the local and global reflectance at each point by a second multi layer perceptron that, in addition, to density features, the current position, the normal (from the signed distace function), view direction, and light position, also takes shadow and highlight hints to aid the network in modeling the corresponding high frequency light transport effects. These hints are provided as a suggestion, and we leave it up to the network to decide how to incorporate these in the final relit result. We demonstrate and validate our neural implicit representation on synthetic and real scenes exhibiting a wide variety of shapes, material properties, and global illumination light transport.
Large language models (LLMs) are capable of performing conditional sequence generation tasks, such as translation or summarization, through instruction fine-tuning. The fine-tuning data is generally sequentially concatenated from a specific task instruction, an input sentence, and the corresponding response. Considering the locality modeled by the self-attention mechanism of LLMs, these models face the risk of instruction forgetting when generating responses for long input sentences. To mitigate this issue, we propose enhancing the instruction-following capability of LLMs by shifting the position of task instructions after the input sentences. Theoretical analysis suggests that our straightforward method can alter the model's learning focus, thereby emphasizing the training of instruction-following capabilities. Concurrently, experimental results demonstrate that our approach consistently outperforms traditional settings across various model scales (1B / 7B / 13B) and different sequence generation tasks (translation and summarization), without any additional data or annotation costs. Notably, our method significantly improves the zero-shot performance on conditional sequence generation, e.g., up to 9.7 BLEU points on WMT zero-shot translation tasks.
The accurate reconstruction of surgical scenes from surgical videos is critical for various applications, including intraoperative navigation and image-guided robotic surgery automation. However, previous approaches, mainly relying on depth estimation, have limited effectiveness in reconstructing surgical scenes with moving surgical tools. To address this limitation and provide accurate 3D position prediction for surgical tools in all frames, we propose a novel approach called SAMSNeRF that combines Segment Anything Model (SAM) and Neural Radiance Field (NeRF) techniques. Our approach generates accurate segmentation masks of surgical tools using SAM, which guides the refinement of the dynamic surgical scene reconstruction by NeRF. Our experimental results on public endoscopy surgical videos demonstrate that our approach successfully reconstructs high-fidelity dynamic surgical scenes and accurately reflects the spatial information of surgical tools. Our proposed approach can significantly enhance surgical navigation and automation by providing surgeons with accurate 3D position information of surgical tools during surgery.The source code will be released soon.