In the context of environmental sound classification, the adaptability of systems is key: which sound classes are interesting depends on the context and the user's needs. Recent advances in text-to-audio retrieval allow for zero-shot audio classification, but performance compared to supervised models remains limited. This work proposes a multimodal prototypical approach that exploits local audio-text embeddings to provide more relevant answers to audio queries, augmenting the adaptability of sound detection in the wild. We do this by first using text to query a nearby community of audio embeddings that best characterize each query sound, and select the group's centroids as our prototypes. Second, we compare unseen audio to these prototypes for classification. We perform multiple ablation studies to understand the impact of the embedding models and prompts. Our unsupervised approach improves upon the zero-shot state-of-the-art in three sound recognition benchmarks by an average of 12%.
相關內容
Pathwise coordinate descent algorithms have been used to compute entire solution paths for lasso and other penalized regression problems quickly with great success. They improve upon cold start algorithms by solving the problems that make up the solution path sequentially for an ordered set of tuning parameter values, instead of solving each problem separately. However, extending pathwise coordinate descent algorithms to more the general bridge or power family of $\ell_q$ penalties is challenging. Faster algorithms for computing solution paths for these penalties are needed because $\ell_q$ penalized regression problems can be nonconvex and especially burdensome to solve. In this paper, we show that a reparameterization of $\ell_q$ penalized regression problems is more amenable to pathwise coordinate descent algorithms. This allows us to improve computation of the mode-thresholding function for $\ell_q$ penalized regression problems in practice and introduce two separate pathwise algorithms. We show that either pathwise algorithm is faster than the corresponding cold-start alternative, and demonstrate that different pathwise algorithms may be more likely to reach better solutions.
We propose a framework that can incrementally expand the explanatory temporal logic rule set to explain the occurrence of temporal events. Leveraging the temporal point process modeling and learning framework, the rule content and weights will be gradually optimized until the likelihood of the observational event sequences is optimal. The proposed algorithm alternates between a master problem, where the current rule set weights are updated, and a subproblem, where a new rule is searched and included to best increase the likelihood. The formulated master problem is convex and relatively easy to solve using continuous optimization, whereas the subproblem requires searching the huge combinatorial rule predicate and relationship space. To tackle this challenge, we propose a neural search policy to learn to generate the new rule content as a sequence of actions. The policy parameters will be trained end-to-end using the reinforcement learning framework, where the reward signals can be efficiently queried by evaluating the subproblem objective. The trained policy can be used to generate new rules in a controllable way. We evaluate our methods on both synthetic and real healthcare datasets, obtaining promising results.
Language models that are sensitive to external context can more effectively capture the speaking patterns of individuals with specific characteristics or in particular environments. However, obtaining and leveraging such annotations can be challenging. In this work, we show how to leverage rich character and film annotations to personalise language models in a scalable manner. Our best model can reduce perplexity by up to 6.5% compared to a parameter-matched language model. Our approach performs on par with speaker-specific fine-tuning when the fine-tuning data (i.e. past dialogue) for individual speakers is available. On top of that, it also generalises well to a scenario with no such data, relying on combinations of demographic characteristics expressed via metadata. Our findings are consistent across two corpora, one of which is also a contribution of this paper: Cornell-rich contains rich manual annotations for 863 speaking characters from the Cornell Movie Dialog Corpus, including features such as characteristic quotes and character descriptions, along with six automatically extracted metadata features for over 95% of the featured films. Finally, we also present a cost-benefit analysis highlighting which annotations are most cost-effective in reducing perplexity.
Hyperproperties extend trace properties to express properties of sets of traces, and they are increasingly popular in specifying various security and performance-related properties in domains such as cyber-physical systems, smart grids, and automotive. This paper introduces a model checking algorithm for a new formalism, HyperTWTL, which extends Time Window Temporal Logic (TWTL) -- a domain-specific formal specification language for robotics, by allowing explicit and simultaneous quantification over multiple execution traces. We present HyperTWTL with both \emph{synchronous} and \emph{asynchronous} semantics, based on the alignment of the timestamps in the traces. Consequently, we demonstrate the application of HyperTWTL in formalizing important information-flow security policies and concurrency for robotics applications. Finally, we propose a model checking algorithm for verifying fragments of HyperTWTL by reducing the problem to a TWTL model checking problem.
Human mesh reconstruction from a single image is challenging in the presence of occlusion, which can be caused by self, objects, or other humans. Existing methods either fail to separate human features accurately or lack proper supervision for feature completion. In this paper, we propose Dense Inpainting Human Mesh Recovery (DIMR), a two-stage method that leverages dense correspondence maps to handle occlusion. Our method utilizes a dense correspondence map to separate visible human features and completes human features on a structured UV map dense human with an attention-based feature completion module. We also design a feature inpainting training procedure that guides the network to learn from unoccluded features. We evaluate our method on several datasets and demonstrate its superior performance under heavily occluded scenarios compared to other methods. Extensive experiments show that our method obviously outperforms prior SOTA methods on heavily occluded images and achieves comparable results on the standard benchmarks (3DPW).
Although audio-visual representation has been proved to be applicable in many downstream tasks, the representation of dancing videos, which is more specific and always accompanied by music with complex auditory contents, remains challenging and uninvestigated. Considering the intrinsic alignment between the cadent movement of dancer and music rhythm, we introduce MuDaR, a novel Music-Dance Representation learning framework to perform the synchronization of music and dance rhythms both in explicit and implicit ways. Specifically, we derive the dance rhythms based on visual appearance and motion cues inspired by the music rhythm analysis. Then the visual rhythms are temporally aligned with the music counterparts, which are extracted by the amplitude of sound intensity. Meanwhile, we exploit the implicit coherence of rhythms implied in audio and visual streams by contrastive learning. The model learns the joint embedding by predicting the temporal consistency between audio-visual pairs. The music-dance representation, together with the capability of detecting audio and visual rhythms, can further be applied to three downstream tasks: (a) dance classification, (b) music-dance retrieval, and (c) music-dance retargeting. Extensive experiments demonstrate that our proposed framework outperforms other self-supervised methods by a large margin.
Humans perceive the world by concurrently processing and fusing high-dimensional inputs from multiple modalities such as vision and audio. Machine perception models, in stark contrast, are typically modality-specific and optimised for unimodal benchmarks, and hence late-stage fusion of final representations or predictions from each modality (`late-fusion') is still a dominant paradigm for multimodal video classification. Instead, we introduce a novel transformer based architecture that uses `fusion bottlenecks' for modality fusion at multiple layers. Compared to traditional pairwise self-attention, our model forces information between different modalities to pass through a small number of bottleneck latents, requiring the model to collate and condense the most relevant information in each modality and only share what is necessary. We find that such a strategy improves fusion performance, at the same time reducing computational cost. We conduct thorough ablation studies, and achieve state-of-the-art results on multiple audio-visual classification benchmarks including Audioset, Epic-Kitchens and VGGSound. All code and models will be released.
Recent contrastive representation learning methods rely on estimating mutual information (MI) between multiple views of an underlying context. E.g., we can derive multiple views of a given image by applying data augmentation, or we can split a sequence into views comprising the past and future of some step in the sequence. Contrastive lower bounds on MI are easy to optimize, but have a strong underestimation bias when estimating large amounts of MI. We propose decomposing the full MI estimation problem into a sum of smaller estimation problems by splitting one of the views into progressively more informed subviews and by applying the chain rule on MI between the decomposed views. This expression contains a sum of unconditional and conditional MI terms, each measuring modest chunks of the total MI, which facilitates approximation via contrastive bounds. To maximize the sum, we formulate a contrastive lower bound on the conditional MI which can be approximated efficiently. We refer to our general approach as Decomposed Estimation of Mutual Information (DEMI). We show that DEMI can capture a larger amount of MI than standard non-decomposed contrastive bounds in a synthetic setting, and learns better representations in a vision domain and for dialogue generation.
Data augmentation has been widely used to improve generalizability of machine learning models. However, comparatively little work studies data augmentation for graphs. This is largely due to the complex, non-Euclidean structure of graphs, which limits possible manipulation operations. Augmentation operations commonly used in vision and language have no analogs for graphs. Our work studies graph data augmentation for graph neural networks (GNNs) in the context of improving semi-supervised node-classification. We discuss practical and theoretical motivations, considerations and strategies for graph data augmentation. Our work shows that neural edge predictors can effectively encode class-homophilic structure to promote intra-class edges and demote inter-class edges in given graph structure, and our main contribution introduces the GAug graph data augmentation framework, which leverages these insights to improve performance in GNN-based node classification via edge prediction. Extensive experiments on multiple benchmarks show that augmentation via GAug improves performance across GNN architectures and datasets.
Benefit from the quick development of deep learning techniques, salient object detection has achieved remarkable progresses recently. However, there still exists following two major challenges that hinder its application in embedded devices, low resolution output and heavy model weight. To this end, this paper presents an accurate yet compact deep network for efficient salient object detection. More specifically, given a coarse saliency prediction in the deepest layer, we first employ residual learning to learn side-output residual features for saliency refinement, which can be achieved with very limited convolutional parameters while keep accuracy. Secondly, we further propose reverse attention to guide such side-output residual learning in a top-down manner. By erasing the current predicted salient regions from side-output features, the network can eventually explore the missing object parts and details which results in high resolution and accuracy. Experiments on six benchmark datasets demonstrate that the proposed approach compares favorably against state-of-the-art methods, and with advantages in terms of simplicity, efficiency (45 FPS) and model size (81 MB).
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191