We propose DiffSHEG, a Diffusion-based approach for Speech-driven Holistic 3D Expression and Gesture generation with arbitrary length. While previous works focused on co-speech gesture or expression generation individually, the joint generation of synchronized expressions and gestures remains barely explored. To address this, our diffusion-based co-speech motion generation transformer enables uni-directional information flow from expression to gesture, facilitating improved matching of joint expression-gesture distributions. Furthermore, we introduce an outpainting-based sampling strategy for arbitrary long sequence generation in diffusion models, offering flexibility and computational efficiency. Our method provides a practical solution that produces high-quality synchronized expression and gesture generation driven by speech. Evaluated on two public datasets, our approach achieves state-of-the-art performance both quantitatively and qualitatively. Additionally, a user study confirms the superiority of DiffSHEG over prior approaches. By enabling the real-time generation of expressive and synchronized motions, DiffSHEG showcases its potential for various applications in the development of digital humans and embodied agents.
Reconfigurable intelligent surfaces (RISs) bring great potential to the advancement of 6G and beyond wireless communication technologies. RISs introduce a great degree of flexibility, allowing some sort of virtual control over the wireless channel. Exploiting the flexibility introduced by RISs, we propose a novel RIS-enabled downlink (DL) non-orthogonal multiple access (NOMA) scheme where NOMA is enabled over-the-air rather than at the base station (BS) or the receiver (Rx). Here, the RIS is partitioned into distinctive groups where each part of the RIS serves a different user equipment (UE) to perform multiple accessing. The BS transmits an unmodulated signal to the RIS, and each partition modulates the impinging signal over-the-air by introducing a phase shift according to the incoming information bits to serve the corresponding UE. First, the end-to-end system model for the proposed system is presented. Furthermore, outage probability calculations, theoretical error probability analysis, and bit error rate (BER) derivations are discussed and reinforced with comprehensive computer simulation results.
We propose UpFusion, a system that can perform novel view synthesis and infer 3D representations for an object given a sparse set of reference images without corresponding pose information. Current sparse-view 3D inference methods typically rely on camera poses to geometrically aggregate information from input views, but are not robust in-the-wild when such information is unavailable/inaccurate. In contrast, UpFusion sidesteps this requirement by learning to implicitly leverage the available images as context in a conditional generative model for synthesizing novel views. We incorporate two complementary forms of conditioning into diffusion models for leveraging the input views: a) via inferring query-view aligned features using a scene-level transformer, b) via intermediate attentional layers that can directly observe the input image tokens. We show that this mechanism allows generating high-fidelity novel views while improving the synthesis quality given additional (unposed) images. We evaluate our approach on the Co3Dv2 and Google Scanned Objects datasets and demonstrate the benefits of our method over pose-reliant sparse-view methods as well as single-view methods that cannot leverage additional views. Finally, we also show that our learned model can generalize beyond the training categories and even allow reconstruction from self-captured images of generic objects in-the-wild.
Most existing debiasing methods for multimodal models, including causal intervention and inference methods, utilize approximate heuristics to represent the biases, such as shallow features from early stages of training or unimodal features for multimodal tasks like VQA, etc., which may not be accurate. In this paper, we study bias arising from confounders in a causal graph for multimodal data and examine a novel approach that leverages causally-motivated information minimization to learn the confounder representations. Robust predictive features contain diverse information that helps a model generalize to out-of-distribution data. Hence, minimizing the information content of features obtained from a pretrained biased model helps learn the simplest predictive features that capture the underlying data distribution. We treat these features as confounder representations and use them via methods motivated by causal theory to remove bias from models. We find that the learned confounder representations indeed capture dataset biases, and the proposed debiasing methods improve out-of-distribution (OOD) performance on multiple multimodal datasets without sacrificing in-distribution performance. Additionally, we introduce a novel metric to quantify the sufficiency of spurious features in models' predictions that further demonstrates the effectiveness of our proposed methods. Our code is available at: https://github.com/Vaidehi99/CausalInfoMin
Existing information retrieval (IR) models often assume a homogeneous format, limiting their applicability to diverse user needs, such as searching for images with text descriptions, searching for a news article with a headline image, or finding a similar photo with a query image. To approach such different information-seeking demands, we introduce UniIR, a unified instruction-guided multimodal retriever capable of handling eight distinct retrieval tasks across modalities. UniIR, a single retrieval system jointly trained on ten diverse multimodal-IR datasets, interprets user instructions to execute various retrieval tasks, demonstrating robust performance across existing datasets and zero-shot generalization to new tasks. Our experiments highlight that multi-task training and instruction tuning are keys to UniIR's generalization ability. Additionally, we construct the M-BEIR, a multimodal retrieval benchmark with comprehensive results, to standardize the evaluation of universal multimodal information retrieval.
We consider a decision aggregation problem with two experts who each make a binary recommendation after observing a private signal about an unknown binary world state. An agent, who does not know the joint information structure between signals and states, sees the experts' recommendations and aims to match the action with the true state. Under the scenario, we study whether supplemented additionally with second-order information (each expert's forecast on the other's recommendation) could enable a better aggregation. We adopt a minimax regret framework to evaluate the aggregator's performance, by comparing it to an omniscient benchmark that knows the joint information structure. With general information structures, we show that second-order information provides no benefit. No aggregator can improve over a trivial aggregator, which always follows the first expert's recommendation. However, positive results emerge when we assume experts' signals are conditionally independent given the world state. When the aggregator is deterministic, we present a robust aggregator that leverages second-order information, which can significantly outperform counterparts without it. Second, when two experts are homogeneous, by adding a non-degenerate assumption on the signals, we demonstrate that random aggregators using second-order information can surpass optimal ones without it. In the remaining settings, the second-order information is not beneficial. We also extend the above results to the setting when the aggregator's utility function is more general.
Many RGBT tracking researches primarily focus on modal fusion design, while overlooking the effective handling of target appearance changes. While some approaches have introduced historical frames or fuse and replace initial templates to incorporate temporal information, they have the risk of disrupting the original target appearance and accumulating errors over time. To alleviate these limitations, we propose a novel Transformer RGBT tracking approach, which mixes spatio-temporal multimodal tokens from the static multimodal templates and multimodal search regions in Transformer to handle target appearance changes, for robust RGBT tracking. We introduce independent dynamic template tokens to interact with the search region, embedding temporal information to address appearance changes, while also retaining the involvement of the initial static template tokens in the joint feature extraction process to ensure the preservation of the original reliable target appearance information that prevent deviations from the target appearance caused by traditional temporal updates. We also use attention mechanisms to enhance the target features of multimodal template tokens by incorporating supplementary modal cues, and make the multimodal search region tokens interact with multimodal dynamic template tokens via attention mechanisms, which facilitates the conveyance of multimodal-enhanced target change information. Our module is inserted into the transformer backbone network and inherits joint feature extraction, search-template matching, and cross-modal interaction. Extensive experiments on three RGBT benchmark datasets show that the proposed approach maintains competitive performance compared to other state-of-the-art tracking algorithms while running at 39.1 FPS.
Data privacy and protection through anonymization is a critical issue for network operators or data owners before it is forwarded for other possible use of data. With the adoption of Artificial Intelligence (AI), data anonymization augments the likelihood of covering up necessary sensitive information; preventing data leakage and information loss. OpenWiFi networks are vulnerable to any adversary who is trying to gain access or knowledge on traffic regardless of the knowledge possessed by data owners. The odds for discovery of actual traffic information is addressed by applied conditional tabular generative adversarial network (CTGAN). CTGAN yields synthetic data; which disguises as actual data but fostering hidden acute information of actual data. In this paper, the similarity assessment of synthetic with actual data is showcased in terms of clustering algorithms followed by a comparison of performance for unsupervised cluster validation metrics. A well-known algorithm, K-means outperforms other algorithms in terms of similarity assessment of synthetic data over real data while achieving nearest scores 0.634, 23714.57, and 0.598 as Silhouette, Calinski and Harabasz and Davies Bouldin metric respectively. On exploiting a comparative analysis in validation scores among several algorithms, K-means forms the epitome of unsupervised clustering algorithms ensuring explicit usage of synthetic data at the same time a replacement for real data. Hence, the experimental results aim to show the viability of using CTGAN-generated synthetic data in lieu of publishing anonymized data to be utilized in various applications.
Autism Spectrum Disorder (ASD) is a complicated neurological condition which is challenging to diagnose. Numerous studies demonstrate that children diagnosed with autism struggle with maintaining attention spans and have less focused vision. The eye-tracking technology has drawn special attention in the context of ASD since anomalies in gaze have long been acknowledged as a defining feature of autism in general. Deep Learning (DL) approaches coupled with eye-tracking sensors are exploiting additional capabilities to advance the diagnostic and its applications. By learning intricate nonlinear input-output relations, DL can accurately recognize the various gaze and eye-tracking patterns and adjust to the data. Convolutions alone are insufficient to capture the important spatial information in gaze patterns or eye tracking. The dynamic kernel-based process known as involutions can improve the efficiency of classifying gaze patterns or eye tracking data. In this paper, we utilise two different image-processing operations to see how these processes learn eye-tracking patterns. Since these patterns are primarily based on spatial information, we use involution with convolution making it a hybrid, which adds location-specific capability to a deep learning model. Our proposed model is implemented in a simple yet effective approach, which makes it easier for applying in real life. We investigate the reasons why our approach works well for classifying eye-tracking patterns. For comparative analysis, we experiment with two separate datasets as well as a combined version of both. The results show that IC with three involution layers outperforms the previous approaches.
In recent years, deep learning has emerged as a potent tool across a multitude of domains, leading to a surge in research pertaining to its application in the wearable human activity recognition (WHAR) domain. Despite the rapid development, concerns have been raised about the lack of standardization and consistency in the procedures used for experimental model training, which may affect the reproducibility and reliability of research results. In this paper, we provide an exhaustive review of contemporary deep learning research in the field of WHAR and collate information pertaining to the training procedure employed in various studies. Our findings suggest that a major trend is the lack of detail provided by model training protocols. Besides, to gain a clearer understanding of the impact of missing descriptions, we utilize a control variables approach to assess the impact of key tunable components (e.g., optimization techniques and early stopping criteria) on the inter-subject generalization capabilities of HAR models. With insights from the analyses, we define a novel integrated training procedure tailored to the WHAR model. Empirical results derived using five well-known \ac{whar} benchmark datasets and three classical HAR model architectures demonstrate the effectiveness of our proposed methodology: in particular, there is a significant improvement in macro F1 leave one subject out cross-validation performance.