This study assesses the performance of a global Local Climate Zone (LCZ) product. We examined the built-type classes of LCZs in three major metropolitan areas within the U.S. A reference LCZ was constructed using a simple rule-based method based on high-resolution 3D building maps. Our evaluation demonstrated that the global LCZ product struggles to differentiate classes that demand precise building footprint information (Classes 6 and 9), and classes that necessitate the identification of subtle differences in building elevation (Classes 4-6). Additionally, we identified inconsistent tendencies, where the distribution of classes skews differently across different cities, suggesting the presence of a data distribution shift problem in the machine learning-based LCZ classifier. Our findings shed light on the uncertainties in global LCZ maps, help identify the LCZ classes that are the most challenging to distinguish, and offer insight into future plans for LCZ development and validation.
Question answering on tabular data (TableQA), which aims at generating answers to questions grounded on a given table, has attracted increasing attention in recent years. Existing work tends to generate factual short-form answers by extracting information from one or a few table cells without reasoning over selected table cells. However, the free-form TableQA, requiring a more complex relevant table cell selection strategy and the complex integration and inference of separate pieces of information, has been under-explored. To this end, this paper proposes a generalized three-stage approach: Table-to-Graph conversion and cell localizing, external knowledge retrieval and table-text fusion (called TAG-QA), addressing the challenge of inferring long free-form answer for generative TableQA. In particular, TAG-QA (1) locates relevant table cells using a graph neural network to gather intersecting cells between relevant rows and columns; (2) leverages external knowledge from Wikipedia and (3) generates answers by integrating both tabular data and natural linguistic information. Experiments with a human evaluation demonstrate that TAG-QA is capable of generating more faithful and coherent sentence when compared with several state-of-the-art baselines. Especially, TAG-QA outperforms the strong pipeline-based baseline TAPAS by 17% and 14%, in terms of BLEU-4 and PARENT F-score, respectively. Moreover, TAG-QA outperforms end-to-end model T5 by 16% and 12% on BLEU-4 and PARENT F-score.
This paper studies contextual biasing with Large Language Models (LLMs), where during second-pass rescoring additional contextual information is provided to a LLM to boost Automatic Speech Recognition (ASR) performance. We propose to leverage prompts for a LLM without fine tuning during rescoring which incorporate a biasing list and few-shot examples to serve as additional information when calculating the score for the hypothesis. In addition to few-shot prompt learning, we propose multi-task training of the LLM to predict both the entity class and the next token. To improve the efficiency for contextual biasing and to avoid exceeding LLMs' maximum sequence lengths, we propose dynamic prompting, where we select the most likely class using the class tag prediction, and only use entities in this class as contexts for next token prediction. Word Error Rate (WER) evaluation is performed on i) an internal calling, messaging, and dictation dataset, and ii) the SLUE-Voxpopuli dataset. Results indicate that biasing lists and few-shot examples can achieve 17.8% and 9.6% relative improvement compared to first pass ASR, and that multi-task training and dynamic prompting can achieve 20.0% and 11.3% relative WER improvement, respectively.
The rapid development of AR/VR brings tremendous demands for 3D content. While the widely-used Computer-Aided Design (CAD) method requires a time-consuming and labor-intensive modeling process, sketch-based 3D modeling offers a potential solution as a natural form of computer-human interaction. However, the sparsity and ambiguity of sketches make it challenging to generate high-fidelity content reflecting creators' ideas. Precise drawing from multiple views or strategic step-by-step drawings is often required to tackle the challenge but is not friendly to novice users. In this work, we introduce a novel end-to-end approach, Deep3DSketch+, which performs 3D modeling using only a single free-hand sketch without inputting multiple sketches or view information. Specifically, we introduce a lightweight generation network for efficient inference in real-time and a structural-aware adversarial training approach with a Stroke Enhancement Module (SEM) to capture the structural information to facilitate learning of the realistic and fine-detailed shape structures for high-fidelity performance. Extensive experiments demonstrated the effectiveness of our approach with the state-of-the-art (SOTA) performance on both synthetic and real datasets.
Panoramic videos contain richer spatial information and have attracted tremendous amounts of attention due to their exceptional experience in some fields such as autonomous driving and virtual reality. However, existing datasets for video segmentation only focus on conventional planar images. To address the challenge, in this paper, we present a panoramic video dataset, PanoVOS. The dataset provides 150 videos with high video resolutions and diverse motions. To quantify the domain gap between 2D planar videos and panoramic videos, we evaluate 15 off-the-shelf video object segmentation (VOS) models on PanoVOS. Through error analysis, we found that all of them fail to tackle pixel-level content discontinues of panoramic videos. Thus, we present a Panoramic Space Consistency Transformer (PSCFormer), which can effectively utilize the semantic boundary information of the previous frame for pixel-level matching with the current frame. Extensive experiments demonstrate that compared with the previous SOTA models, our PSCFormer network exhibits a great advantage in terms of segmentation results under the panoramic setting. Our dataset poses new challenges in panoramic VOS and we hope that our PanoVOS can advance the development of panoramic segmentation/tracking.
In, Elmarakeby et al., "Biologically informed deep neural network for prostate cancer discovery", a feedforward neural network with biologically informed, sparse connections (P-NET) was presented to model the state of prostate cancer. We verified the reproducibility of the study conducted by Elmarakeby et al., using both their original codebase, and our own re-implementation using more up-to-date libraries. We quantified the contribution of network sparsification by Reactome biological pathways, and confirmed its importance to P-NET's superior performance. Furthermore, we explored alternative neural architectures and approaches to incorporating biological information into the networks. We experimented with three types of graph neural networks on the same training data, and investigated the clinical prediction agreement between different models. Our analyses demonstrated that deep neural networks with distinct architectures make incorrect predictions for individual patient that are persistent across different initializations of a specific neural architecture. This suggests that different neural architectures are sensitive to different aspects of the data, an important yet under-explored challenge for clinical prediction tasks.
In the past decades, a variety of haptic devices have been developed to facilitate high-fidelity human-computer interaction (HCI) in virtual reality (VR). In particular, passive haptic feedback can create a compelling sensation based on real objects spatially overlapping with their virtual counterparts. However, these approaches require pre-deployment efforts, hindering their democratizing use in practice. We propose the Tracking Calibrated Robot (TCR), a novel and general haptic approach to free developers from deployment efforts, which can be potentially deployed in any scenario. Specifically, we augment the VR with a collaborative robot that renders haptic contact in the real world while the user touches a virtual object in the virtual world. The distance between the user's finger and the robot end-effector is controlled over time. The distance starts to smoothly reduce to zero when the user intends to touch the virtual object. A mock user study tested users' perception of three virtual objects, and the result shows that TCR is effective in terms of conveying discriminative shape information.
Successful out-of-distribution generalization requires environment annotations. Unfortunately, these are resource-intensive to obtain, and their relevance to model performance is limited by the expectations and perceptual biases of human annotators. Therefore, to enable robust AI systems across applications, we must develop algorithms to automatically discover environments inducing broad generalization. Current proposals, which divide examples based on their training error, suffer from one fundamental problem. These methods add hyper-parameters and early-stopping criteria that are impossible to tune without a validation set with human-annotated environments, the very information subject to discovery. In this paper, we propose Cross-Risk-Minimization (XRM) to address this issue. XRM trains two twin networks, each learning from one random half of the training data, while imitating confident held-out mistakes made by its sibling. XRM provides a recipe for hyper-parameter tuning, does not require early-stopping, and can discover environments for all training and validation data. Domain generalization algorithms built on top of XRM environments achieve oracle worst-group-accuracy, solving a long-standing problem in out-of-distribution generalization.
Offline Reinforcement Learning (RL) has emerged as a promising framework for learning policies without active interactions, making it especially appealing for autonomous driving tasks. Recent successes of Transformers inspire casting offline RL as sequence modeling, which performs well in long-horizon tasks. However, they are overly optimistic in stochastic environments with incorrect assumptions that the same goal can be consistently achieved by identical actions. In this paper, we introduce an UNcertainty-awaRE deciSion Transformer (UNREST) for planning in stochastic driving environments without introducing additional transition or complex generative models. Specifically, UNREST estimates state uncertainties by the conditional mutual information between transitions and returns, and segments sequences accordingly. Discovering the `uncertainty accumulation' and `temporal locality' properties of driving environments, UNREST replaces the global returns in decision transformers with less uncertain truncated returns, to learn from true outcomes of agent actions rather than environment transitions. We also dynamically evaluate environmental uncertainty during inference for cautious planning. Extensive experimental results demonstrate UNREST's superior performance in various driving scenarios and the power of our uncertainty estimation strategy.
Existing 3D model reconstruction methods typically produce outputs in the form of voxels, point clouds, or meshes. However, each of these approaches has its limitations and may not be suitable for every scenario. For instance, the resulting model may exhibit a rough surface and distorted structure, making manual editing and post-processing challenging for humans. In this paper, we introduce a novel 3D reconstruction method designed to address these issues. We trained a visual transformer to predict a "scene descriptor" from a single wire-frame image. This descriptor encompasses crucial information, including object types and parameters such as position, rotation, and size. With the predicted parameters, a 3D scene can be reconstructed using 3D modeling software like Blender or Rhino Grasshopper which provides a programmable interface, resulting in finely and easily editable 3D models. To evaluate the proposed model, we created two datasets: one featuring simple scenes and another with complex scenes. The test results demonstrate the model's ability to accurately reconstruct simple scenes but reveal its challenges with more complex ones.