Despite the notable advancements in numerous Transformer-based models, the task of long multi-horizon time series forecasting remains a persistent challenge, especially towards explainability. Focusing on commonly used saliency maps in explaining DNN in general, our quest is to build attention-based architecture that can automatically encode saliency-related temporal patterns by establishing connections with appropriate attention heads. Hence, this paper introduces Temporal Saliency Detection (TSD), an effective approach that builds upon the attention mechanism and applies it to multi-horizon time series prediction. While our proposed architecture adheres to the general encoder-decoder structure, it undergoes a significant renovation in the encoder component, wherein we incorporate a series of information contracting and expanding blocks inspired by the U-Net style architecture. The TSD approach facilitates the multiresolution analysis of saliency patterns by condensing multi-heads, thereby progressively enhancing the forecasting of complex time series data. Empirical evaluations illustrate the superiority of our proposed approach compared to other models across multiple standard benchmark datasets in diverse far-horizon forecasting settings. The initial TSD achieves substantial relative improvements of 31% and 46% over several models in the context of multivariate and univariate prediction. We believe the comprehensive investigations presented in this study will offer valuable insights and benefits to future research endeavors.
相關內容
This paper addresses the problem of anticipating the next-active-object location in the future, for a given egocentric video clip where the contact might happen, before any action takes place. The problem is considerably hard, as we aim at estimating the position of such objects in a scenario where the observed clip and the action segment are separated by the so-called ``time to contact'' (TTC) segment. Many methods have been proposed to anticipate the action of a person based on previous hand movements and interactions with the surroundings. However, there have been no attempts to investigate the next possible interactable object, and its future location with respect to the first-person's motion and the field-of-view drift during the TTC window. We define this as the task of Anticipating the Next ACTive Object (ANACTO). To this end, we propose a transformer-based self-attention framework to identify and locate the next-active-object in an egocentric clip. We benchmark our method on three datasets: EpicKitchens-100, EGTEA+ and Ego4D. We also provide annotations for the first two datasets. Our approach performs best compared to relevant baseline methods. We also conduct ablation studies to understand the effectiveness of the proposed and baseline methods on varying conditions. Code and ANACTO task annotations will be made available upon paper acceptance.
Remarkable advancements have been made recently in point cloud analysis through the exploration of transformer architecture, but it remains challenging to effectively learn local and global structures within point clouds. In this paper, we propose a new transformer network equipped with a collect-and-distribute mechanism to communicate short- and long-range contexts of point clouds, which we refer to as CDFormer. Specifically, we first employ self-attention to capture short-range interactions within each local patch, and the updated local features are then collected into a set of proxy reference points from which we can extract long-range contexts. Afterward, we distribute the learned long-range contexts back to local points via cross-attention. To address the position clues for short- and long-range contexts, we additionally introduce the context-aware position encoding to facilitate position-aware communications between points. We perform experiments on five popular point cloud datasets, namely ModelNet40, ScanObjectNN, ShapeNetPart, S3DIS and ScanNetV2, for classification and segmentation. Results show the effectiveness of the proposed CDFormer, delivering several new state-of-the-art performances on point cloud classification and segmentation tasks. The source code is available at \url{//github.com/haibo-qiu/CDFormer}.
Accurate modeling of the diverse and dynamic interests of users remains a significant challenge in the design of personalized recommender systems. Existing user modeling methods, like single-point and multi-point representations, have limitations w.r.t. accuracy, diversity, computational cost, and adaptability. To overcome these deficiencies, we introduce density-based user representations (DURs), a novel model that leverages Gaussian process regression for effective multi-interest recommendation and retrieval. Our approach, GPR4DUR, exploits DURs to capture user interest variability without manual tuning, incorporates uncertainty-awareness, and scales well to large numbers of users. Experiments using real-world offline datasets confirm the adaptability and efficiency of GPR4DUR, while online experiments with simulated users demonstrate its ability to address the exploration-exploitation trade-off by effectively utilizing model uncertainty.
In this work, a cut high-dimensional model representation (cut-HDMR) expansion based on multiple anchors is constructed via the clustering method. Specifically, a set of random input realizations is drawn from the parameter space and grouped by the centroidal Voronoi tessellation (CVT) method. Then for each cluster, the centroid is set as the reference, thereby the corresponding zeroth-order term can be determined directly. While for non-zero order terms of each cut-HDMR, a set of discrete points is selected for each input component, and the Lagrange interpolation method is applied. For a new input, the cut-HDMR corresponding to the nearest centroid is used to compute its response. Numerical experiments with high-dimensional integral and elliptic stochastic partial differential equation as backgrounds show that the CVT based multiple anchors cut-HDMR can alleviate the negative impact of a single inappropriate anchor point, and has higher accuracy than the average of several expansions.
Large language models (LLMs) struggle on processing complicated observations in interactive decision making tasks. To alleviate this issue, we propose a simple hierarchical prompting approach. Diverging from previous prompting approaches that always put the full observation (e.g. a web page) to the prompt, we propose to first construct an action-aware observation which is more condensed and relevant with a dedicated SUMMARIZER prompt. The ACTOR prompt then predicts the next action based on the summarized observation. While our method has broad applicability, we particularly demonstrate its efficacy in the complex domain of web navigation where a full observation often contains redundant and irrelevant information. Our approach outperforms the previous state-of-the-art prompting mechanics by 6.2% on task success rate, demonstrating its potential on interactive decision making tasks with long observation traces.
Path planners based on basic rapidly-exploring random trees (RRTs) are quick and efficient, and thus favourable for real-time robot path planning, but are almost-surely suboptimal. In contrast, the optimal RRT (RRT*) converges to the optimal solution, but may be expensive in practice. Recent work has focused on accelerating the RRT*'s convergence rate. The most successful strategies are informed sampling, path optimisation, and a combination thereof. However, informed sampling and its combination with path optimisation have not been applied to the basic RRT. Moreover, while a number of path optimisers can be used to accelerate the convergence rate, a comparison of their effectiveness is lacking. This paper investigates the use of informed sampling and path optimisation to accelerate planners based on both the basic RRT and the RRT*, resulting in a family of algorithms known as optimised informed RRTs. We apply different path optimisers and compare their effectiveness. The goal is to ascertain if applying informed sampling and path optimisation can help the quick, though almost-surely suboptimal, path planners based on the basic RRT attain comparable or better performance than RRT*-based planners. Analyses show that RRT-based optimised informed RRTs do attain better performance than their RRT*-based counterparts, both when planning time is limited and when there is more planning time.
CLIP, one of the pioneering foundation models that connect images and text, has enabled many recent breakthroughs in computer vision. However, its associated training cost is prohibitively high, imposing a significant barrier to its widespread exploration. In this paper, we present a surprising finding that there exists an inverse scaling law for CLIP training, whereby the larger the image/text encoders used, the shorter the sequence length of image/text tokens that can be applied in training. Moreover, we showcase that the strategy for reducing image/text token length plays a crucial role in determining the quality of this scaling law. As a result of this finding, we are able to successfully train CLIP even with limited computational resources. For example, using 8 A100 GPUs, our CLIP models achieve zero-shot top-1 ImageNet-1k accuracies of 63.2% in ~2 days, 67.8% in ~3 days, and 69.3% in ~4 days. Our method also works well when scaling up -- with G/14, we register a new record of 83.0% ImageNet-1k zero-shot accuracy, and meanwhile accelerate the training by ~33x compared to its OpenCLIP counterpart. By reducing the computation barrier associated with CLIP, we hope to inspire more research in this field, particularly from academics. Our code is available at //github.com/UCSC-VLAA/CLIPA.
The metaverse refers to the merger of technologies for providing a digital twin of the real world and the underlying connectivity and interactions for the many kinds of agents within. As this set of technology paradigms - involving artificial intelligence, mixed reality, the internet-of-things and others - gains in scale, maturity, and utility there are rapidly emerging design challenges and new research opportunities. In particular is the metaverse disconnect problem, the gap in task switching that inevitably occurs when a user engages with multiple virtual and physical environments simultaneously. Addressing this gap remains an open issue that affects the user experience and must be overcome to increase overall utility of the metaverse. This article presents design frameworks that consider how to address the metaverse as a hyper-connected meta-environment that connects and expands multiple user environments, modalities, contexts, and the many objects and relationships within them. This article contributes to i) a framing of the metaverse as a social XR-IoT (XRI) concept, ii) design Considerations for XRI metaverse experiences, iii) a design architecture for social multi-user XRI metaverse environments, and iv) descriptive exploration of social interaction scenarios within XRI multi-user metaverses. These contribute a new design framework for metaverse researchers and creators to consider the coming wave of interconnected and immersive smart environments.
Simultaneous sequence generation is a pivotal task for real-time scenarios, such as streaming speech recognition, simultaneous machine translation and simultaneous speech translation, where the target sequence is generated while receiving the source sequence. The crux of achieving high-quality generation with low latency lies in identifying the optimal moments for generating, accomplished by learning a mapping between the source and target sequences. However, existing methods often rely on task-specific heuristics for different sequence types, limiting the model's capacity to adaptively learn the source-target mapping and hindering the exploration of multi-task learning for various simultaneous tasks. In this paper, we propose a unified segment-to-segment framework (Seg2Seg) for simultaneous sequence generation, which learns the mapping in an adaptive and unified manner. During the process of simultaneous generation, the model alternates between waiting for a source segment and generating a target segment, making the segment serve as the natural bridge between the source and target. To accomplish this, Seg2Seg introduces a latent segment as the pivot between source to target and explores all potential source-target mappings via the proposed expectation training, thereby learning the optimal moments for generating. Experiments on multiple simultaneous generation tasks demonstrate that Seg2Seg achieves state-of-the-art performance and exhibits better generality across various tasks.
Recent advancements in deep neural networks for graph-structured data have led to state-of-the-art performance on recommender system benchmarks. However, making these methods practical and scalable to web-scale recommendation tasks with billions of items and hundreds of millions of users remains a challenge. Here we describe a large-scale deep recommendation engine that we developed and deployed at Pinterest. We develop a data-efficient Graph Convolutional Network (GCN) algorithm PinSage, which combines efficient random walks and graph convolutions to generate embeddings of nodes (i.e., items) that incorporate both graph structure as well as node feature information. Compared to prior GCN approaches, we develop a novel method based on highly efficient random walks to structure the convolutions and design a novel training strategy that relies on harder-and-harder training examples to improve robustness and convergence of the model. We also develop an efficient MapReduce model inference algorithm to generate embeddings using a trained model. We deploy PinSage at Pinterest and train it on 7.5 billion examples on a graph with 3 billion nodes representing pins and boards, and 18 billion edges. According to offline metrics, user studies and A/B tests, PinSage generates higher-quality recommendations than comparable deep learning and graph-based alternatives. To our knowledge, this is the largest application of deep graph embeddings to date and paves the way for a new generation of web-scale recommender systems based on graph convolutional architectures.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191