Neural radiance fields (NeRF) show great success in novel view synthesis. However, in real-world scenes, recovering high-quality details from the source images is still challenging for the existing NeRF-based approaches, due to the potential imperfect calibration information and scene representation inaccuracy. Even with high-quality training frames, the synthetic novel views produced by NeRF models still suffer from notable rendering artifacts, such as noise, blur, etc. Towards to improve the synthesis quality of NeRF-based approaches, we propose NeRFLiX, a general NeRF-agnostic restorer paradigm by learning a degradation-driven inter-viewpoint mixer. Specially, we design a NeRF-style degradation modeling approach and construct large-scale training data, enabling the possibility of effectively removing NeRF-native rendering artifacts for existing deep neural networks. Moreover, beyond the degradation removal, we propose an inter-viewpoint aggregation framework that is able to fuse highly related high-quality training images, pushing the performance of cutting-edge NeRF models to entirely new levels and producing highly photo-realistic synthetic views.
Seeing only a tiny part of the whole is not knowing the full circumstance. Bird's-eye-view (BEV) perception, a process of obtaining allocentric maps from egocentric views, is restricted when using a narrow Field of View (FoV) alone. In this work, mapping from 360{\deg} panoramas to BEV semantics, the 360BEV task, is established for the first time to achieve holistic representations of indoor scenes in a top-down view. Instead of relying on narrow-FoV image sequences, a panoramic image with depth information is sufficient to generate a holistic BEV semantic map. To benchmark 360BEV, we present two indoor datasets, 360BEV-Matterport and 360BEV-Stanford, both of which include egocentric panoramic images and semantic segmentation labels, as well as allocentric semantic maps. Besides delving deep into different mapping paradigms, we propose a dedicated solution for panoramic semantic mapping, namely 360Mapper. Through extensive experiments, our methods achieve 44.32% and 45.78% in mIoU on both datasets respectively, surpassing previous counterparts with gains of +7.60% and +9.70% in mIoU. Code and datasets will be available at: https://jamycheung.github.io/360BEV.html.
Confidence-based pseudo-labeling is among the dominant approaches in semi-supervised learning (SSL). It relies on including high-confidence predictions made on unlabeled data as additional targets to train the model. We propose ProtoCon, a novel SSL method aimed at the less-explored label-scarce SSL where such methods usually underperform. ProtoCon refines the pseudo-labels by leveraging their nearest neighbours' information. The neighbours are identified as the training proceeds using an online clustering approach operating in an embedding space trained via a prototypical loss to encourage well-formed clusters. The online nature of ProtoCon allows it to utilise the label history of the entire dataset in one training cycle to refine labels in the following cycle without the need to store image embeddings. Hence, it can seamlessly scale to larger datasets at a low cost. Finally, ProtoCon addresses the poor training signal in the initial phase of training (due to fewer confident predictions) by introducing an auxiliary self-supervised loss. It delivers significant gains and faster convergence over state-of-the-art across 5 datasets, including CIFARs, ImageNet and DomainNet.
Predicting the next location is a highly valuable and common need in many location-based services such as destination prediction and route planning. The goal of next location recommendation is to predict the next point-of-interest a user might go to based on the user's historical trajectory. Most existing models learn mobility patterns merely from users' historical check-in sequences while overlooking the significance of user preference modeling. In this work, a novel Point-of-Interest Transformer (POIFormer) with contrastive user preference modeling is developed for end-to-end next location recommendation. This model consists of three major modules: history encoder, query generator, and preference decoder. History encoder is designed to model mobility patterns from historical check-in sequences, while query generator explicitly learns user preferences to generate user-specific intention queries. Finally, preference decoder combines the intention queries and historical information to predict the user's next location. Extensive comparisons with representative schemes and ablation studies on four real-world datasets demonstrate the effectiveness and superiority of the proposed scheme under various settings.
In orthogonal time frequency space (OTFS) modulation, Zak transform approach is a natural approach for converting information symbols multiplexed in the DD domain directly to time domain for transmission, and vice versa at the receiver. Past research on OTFS has primarily considered a two-step approach where DD domain symbols are first converted to time-frequency domain which are then converted to time domain for transmission, and vice versa at the receiver. The Zak transform approach can offer performance and complexity benefits compared to the two-step approach. This paper presents an early investigation on the bit error performance of OTFS realized using discrete Zak transform (DZT). We develop a compact DD domain input-output relation for DZT-OTFS using matrix decomposition that is valid for both integer and fractional delay-Dopplers. We analyze the bit error performance of DZT-OTFS using pairwise error probability analysis and simulations. Simulation results show that 1) both DZT-OTFS and two-step OTFS perform better than OFDM, and 2) DZT-OTFS achieves better performance compared to two-step OTFS over a wide range of Doppler spreads.
Quantum process learning is emerging as an important tool to study quantum systems. While studied extensively in coherent frameworks, where the target and model system can share quantum information, less attention has been paid to whether the dynamics of quantum systems can be learned without the system and target directly interacting. Such incoherent frameworks are practically appealing since they open up methods of transpiling quantum processes between the different physical platforms without the need for technically challenging hybrid entanglement schemes. Here we provide bounds on the sample complexity of learning unitary processes incoherently by analyzing the number of measurements that are required to emulate well-established coherent learning strategies. We prove that if arbitrary measurements are allowed, then any efficiently representable unitary can be efficiently learned within the incoherent framework; however, when restricted to shallow-depth measurements only low-entangling unitaries can be learned. We demonstrate our incoherent learning algorithm for low entangling unitaries by successfully learning a 16-qubit unitary on \texttt{ibmq\_kolkata}, and further demonstrate the scalabilty of our proposed algorithm through extensive numerical experiments.
Inertia Axes are involved in many techniques for image content measurement when involving information obtained from lines, angles, centroids... etc. We investigate, here, the estimation of the main axis of inertia of an object in the image. We identify the coincidence conditions of the Scale Space Radon Transform (SSRT) maximum and the inertia main axis. We show, that by choosing the appropriate scale parameter, it is possible to match the SSRT maximum and the main axis of inertia location and orientation of the embedded object in the image. Furthermore, an example of use case is presented where binary objects central symmetry computation is derived by means of SSRT projections and the axis of inertia orientation. To this end, some SSRT characteristics have been highlighted and exploited. The experimentations show the SSRT-based main axis of inertia computation effectiveness. Concerning the central symmetry, results are very satisfying as experimentations carried out on randomly created images dataset and existing datasets have permitted to divide successfully these images bases into centrally symmetric and non-centrally symmetric objects.
We introduce a new open information extraction (OIE) benchmark for pre-trained language models (LM). Recent studies have demonstrated that pre-trained LMs, such as BERT and GPT, may store linguistic and relational knowledge. In particular, LMs are able to answer ``fill-in-the-blank'' questions when given a pre-defined relation category. Instead of focusing on pre-defined relations, we create an OIE benchmark aiming to fully examine the open relational information present in the pre-trained LMs. We accomplish this by turning pre-trained LMs into zero-shot OIE systems. Surprisingly, pre-trained LMs are able to obtain competitive performance on both standard OIE datasets (CaRB and Re-OIE2016) and two new large-scale factual OIE datasets (TAC KBP-OIE and Wikidata-OIE) that we establish via distant supervision. For instance, the zero-shot pre-trained LMs outperform the F1 score of the state-of-the-art supervised OIE methods on our factual OIE datasets without needing to use any training sets. Our code and datasets are available at https://github.com/cgraywang/IELM
Recently, a large number of studies have shown that the introduction of visual information can effectively improve the effect of neural machine translation (NMT). Its effectiveness largely depends on the availability of a large number of bilingual parallel sentence pairs and manual image annotation. The lack of images and the effectiveness of images have been difficult to solve. In this paper, a multimodal pre-training generalization algorithm for self-supervised training is proposed, which overcomes the lack of visual information and inaccuracy, and thus extends the applicability of images on NMT. Specifically, we will search for many pictures from the existing sentences through the search engine, and then through the relationship between visual information and text, do the self-supervised training task of graphics and text to obtain more effective visual information for text. We show that when the filtered information is used as multimodal machine translation for fine-tuning, the effect of translation in the global voice dataset is 0.5 BLEU higher than the baseline.
We propose a novel type of map for visual navigation, a renderable neural radiance map (RNR-Map), which is designed to contain the overall visual information of a 3D environment. The RNR-Map has a grid form and consists of latent codes at each pixel. These latent codes are embedded from image observations, and can be converted to the neural radiance field which enables image rendering given a camera pose. The recorded latent codes implicitly contain visual information about the environment, which makes the RNR-Map visually descriptive. This visual information in RNR-Map can be a useful guideline for visual localization and navigation. We develop localization and navigation frameworks that can effectively utilize the RNR-Map. We evaluate the proposed frameworks on camera tracking, visual localization, and image-goal navigation. Experimental results show that the RNR-Map-based localization framework can find the target location based on a single query image with fast speed and competitive accuracy compared to other baselines. Also, this localization framework is robust to environmental changes, and even finds the most visually similar places when a query image from a different environment is given. The proposed navigation framework outperforms the existing image-goal navigation methods in difficult scenarios, under odometry and actuation noises. The navigation framework shows 65.7% success rate in curved scenarios of the NRNS dataset, which is an improvement of 18.6% over the current state-of-the-art.