Text-video retrieval aims to find the most relevant cross-modal samples for a given query. Recent methods focus on modeling the whole spatial-temporal relations. However, since video clips contain more diverse content than captions, the model aligning these asymmetric video-text pairs has a high risk of retrieving many false positive results. In this paper, we propose Probabilistic Token Aggregation (\textit{ProTA}) to handle cross-modal interaction with content asymmetry. Specifically, we propose dual partial-related aggregation to disentangle and re-aggregate token representations in both low-dimension and high-dimension spaces. We propose token-based probabilistic alignment to generate token-level probabilistic representation and maintain the feature representation diversity. In addition, an adaptive contrastive loss is proposed to learn compact cross-modal distribution space. Based on extensive experiments, \textit{ProTA} achieves significant improvements on MSR-VTT (50.9%), LSMDC (25.8%), and DiDeMo (47.2%).
This paper explores an energy-modified leverage sampling strategy for matrix completion in radio map construction. The main goal is to address potential identifiability issues in matrix completion with sparse observations by using a probabilistic sampling approach. Although conventional leverage sampling is commonly employed for designing sampling patterns, it often assigns high sampling probability to locations with low received signal strength (RSS) values, leading to a low sampling efficiency. Theoretical analysis demonstrates that the leverage score produces pseudo images of sources, and in the regions around the source locations, the leverage probability is asymptotically consistent with the RSS. Based on this finding, an energy-modified leverage probability-based sampling strategy is investigated for efficient sampling. Numerical demonstrations indicate that the proposed sampling strategy can decrease the normalized mean squared error (NMSE) of radio map construction by more than 10% for both matrix completion and interpolation-assisted matrix completion schemes, compared to conventional methods.
Few-shot OOD detection focuses on recognizing out-of-distribution (OOD) images that belong to classes unseen during training, with the use of only a small number of labeled in-distribution (ID) images. Up to now, a mainstream strategy is based on large-scale vision-language models, such as CLIP. However, these methods overlook a crucial issue: the lack of reliable OOD supervision information, which can lead to biased boundaries between in-distribution (ID) and OOD. To tackle this problem, we propose CLIP-driven Outliers Synthesis~(CLIP-OS). Firstly, CLIP-OS enhances patch-level features' perception by newly proposed patch uniform convolution, and adaptively obtains the proportion of ID-relevant information by employing CLIP-surgery-discrepancy, thus achieving separation between ID-relevant and ID-irrelevant. Next, CLIP-OS synthesizes reliable OOD data by mixing up ID-relevant features from different classes to provide OOD supervision information. Afterward, CLIP-OS leverages synthetic OOD samples by unknown-aware prompt learning to enhance the separability of ID and OOD. Extensive experiments across multiple benchmarks demonstrate that CLIP-OS achieves superior few-shot OOD detection capability.
Object-centric learning aims to break down complex visual scenes into more manageable object representations, enhancing the understanding and reasoning abilities of machine learning systems toward the physical world. Recently, slot-based video models have demonstrated remarkable proficiency in segmenting and tracking objects, but they overlook the importance of the effective reasoning module. In the real world, reasoning and predictive abilities play a crucial role in human perception and object tracking; in particular, these abilities are closely related to human intuitive physics. Inspired by this, we designed a novel reasoning module called the Slot-based Time-Space Transformer with Memory buffer (STATM) to enhance the model's perception ability in complex scenes. The memory buffer primarily serves as storage for slot information from upstream modules, the Slot-based Time-Space Transformer makes predictions through slot-based spatiotemporal attention computations and fusion. Our experiment results on various datasets show that STATM can significantly enhance object-centric learning capabilities of slot-based video models.
Neural radiance field has achieved fundamental success in novel view synthesis from input views with the same brightness level captured under fixed normal lighting. Unfortunately, synthesizing novel views remains to be a challenge for input views with heterogeneous brightness level captured under low-light condition. The condition is pretty common in the real world. It causes low-contrast images where details are concealed in the darkness and camera sensor noise significantly degrades the image quality. To tackle this problem, we propose to learn to decompose illumination, reflectance, and noise from input views according to that reflectance remains invariant across heterogeneous views. To cope with heterogeneous brightness and noise levels across multi-views, we learn an illumination embedding and optimize a noise map individually for each view. To allow intuitive editing of the illumination, we design an illumination adjustment module to enable either brightening or darkening of the illumination component. Comprehensive experiments demonstrate that this approach enables effective intrinsic decomposition for low-light multi-view noisy images and achieves superior visual quality and numerical performance for synthesizing novel views compared to state-of-the-art methods.
The prevailing approach to aligning Large Language Models (LLMs) typically relies on human or AI feedback and assumes access to specific types of preference datasets. In our work, we question the efficacy of such datasets and explore various scenarios where alignment with expert demonstrations proves more realistic. We build a sequential decision-making framework to formulate the problem of aligning LLMs using demonstration datasets. Drawing insights from inverse reinforcement learning and imitation learning, we introduce various approaches for divergence minimization in the LLM alignment tasks. Our analysis highlights the mass-covering and mode-seeking behaviors of these different approaches. Inclusively, we examine the pros and cons of the classical supervised fine-tuning method, elaborating on scenarios where different methods shine.
Chinese Spell Checking (CSC) is a widely used technology, which plays a vital role in speech to text (STT) and optical character recognition (OCR). Most of the existing CSC approaches relying on BERT architecture achieve excellent performance. However, limited by the scale of the foundation model, BERT-based method does not work well in few-shot scenarios, showing certain limitations in practical applications. In this paper, we explore using an in-context learning method named RS-LLM (Rich Semantic based LLMs) to introduce large language models (LLMs) as the foundation model. Besides, we study the impact of introducing various Chinese rich semantic information in our framework. We found that by introducing a small number of specific Chinese rich semantic structures, LLMs achieve better performance than the BERT-based model on few-shot CSC task. Furthermore, we conduct experiments on multiple datasets, and the experimental results verified the superiority of our proposed framework.
Current disfluency detection methods heavily rely on costly and scarce human-annotated data. To tackle this issue, some approaches employ heuristic or statistical features to generate disfluent sentences, partially improving detection performance. However, these sentences often deviate from real-life scenarios, constraining overall model enhancement. In this study, we propose a lightweight data augmentation approach for disfluency detection, utilizing the superior generative and semantic understanding capabilities of large language model (LLM) to generate disfluent sentences as augmentation data. We leverage LLM to generate diverse and more realistic sentences guided by specific prompts, without the need for fine-tuning the LLM. Subsequently, we apply an uncertainty-aware data filtering approach to improve the quality of the generated sentences, utilized in training a small detection model for improved performance. Experiments using enhanced data yielded state-of-the-art results. The results showed that using a small amount of LLM-generated enhanced data can significantly improve performance, thereby further enhancing cost-effectiveness.
The remarkable capability of large language models (LLMs) for in-context learning (ICL) needs to be activated by demonstration examples. Prior work has extensively explored the selection of examples for ICL, predominantly following the "select then organize" paradigm, such approaches often neglect the internal relationships between examples and exist an inconsistency between the training and inference. In this paper, we formulate the problem as a $\textit{se}$quential $\textit{se}$lection problem and introduce $Se^2$, a sequential-aware method that leverages the LLM's feedback on varying context, aiding in capturing inter-relationships and sequential information among examples, significantly enriching the contextuality and relevance of ICL prompts. Meanwhile, we utilize beam search to seek and construct example sequences, enhancing both quality and diversity. Extensive experiments across 23 NLP tasks from 8 distinct categories illustrate that $Se^2$ markedly surpasses competitive baselines and achieves 42% relative improvement over random selection. Further in-depth analysis show the effectiveness of proposed strategies, highlighting $Se^2$'s exceptional stability and adaptability across various scenarios. Our code will be released to facilitate future research.
There is a growing interest in utilizing machine learning (ML) methods for structural metamodeling due to the substantial computational cost of traditional numerical simulations. The existing data-driven strategies show potential limitations to the model robustness and interpretability as well as the dependency of rich data. To address these challenges, this paper presents a novel physics-informed machine learning (PiML) method, which incorporates scientific principles and physical laws into deep neural networks for modeling seismic responses of nonlinear structures. The basic concept is to constrain the solution space of the ML model within known physical bounds. This is made possible with three main features, namely, model order reduction, a long short-term memory (LSTM) networks, and Newton's second law (e.g., the equation of motion). Model order reduction is essential for handling structural systems with inherent redundancy and enhancing model efficiency. The LSTM network captures temporal dependencies, enabling accurate prediction of time series responses. The equation of motion is manipulated to learn system nonlinearities and confines the solution space within physically interpretable results. These features enable model training with relatively sparse data and offer benefits in terms of accuracy, interpretability, and robustness. Furthermore, a dataset of seismically designed archetype ductile planar steel moment resistant frames under horizontal seismic loading, available in the DesignSafe-CI Database, is considered for evaluation of the proposed method. The resulting metamodel is capable of handling more complex data compared to existing physics-guided LSTM models and outperforms other non-physics data-driven neural networks.