Audio-to-image generation (AIG) faces challenges in fine-grained semantic alignment, particularly semantic semantic misalignment, and loss of visual detail. To address these issues, we proposed MARS ( M ultimodal- A ssisted R efined S emantic alignment), a novel framework leveraging a Mamba-based audio encoder to manage the complexity of long audio sequences, coupled with a fine-grained multimodal alignment strategy using visual descriptions from multimodal large language models. We enhanced semantic coherence and aesthetic quality by fine-tuning an image generator using an image aesthetic perception generator. Furthermore, we validated MARS on VGGSound and VEGAS benchmarks, comprising 37,250 and 9,500 records, respectively. The results suggest that MARS significantly outperforms existing methods, achieving average improvements of 28.73% in semantic relevance and 127.35% in aesthetic scores compared with the best AIG generation baseline. In addition, cross-domain evaluations on the AudioCaps and Clotho datasets confirmed the robustness and generalization capability of MARS , with an average improvement of 73.9% on the V2A metric. • MARS refines semantic alignment for audio-based image generation. • A Mamba-based encoder processes long audio sequences. • MLLMs supply rich visuals to boost semantics and aesthetics. • Extensive experiments prove effectiveness and robustness.

With the introduction of large language models (LLMs), automatic math reasoning has seen tremendous success. However, current methods primarily focus on providing solutions or using techniques like Chain-of-Thought to enhance problem-solving accuracy. In this paper, we focus on improving the capability of mathematics teaching via a Socratic teaching-based LLM (SocraticLLM), which guides learners toward profound thinking with clarity and self-discovery via conversation. We collect and release a high-quality mathematical teaching dataset, named SocraticMATH, which provides Socratic-style conversations of problems with extra knowledge. Also, we propose a knowledge-enhanced LLM as a strong baseline to generate reliable responses with review, guidance/heuristic, rectification, and summarization. Experimental results show the great advantages of SocraticLLM by comparing it with several strong generative models. The codes and datasets are available on https://github.com/ECNU-ICALK/SocraticMath.

Retrieval-Augmented Generation (RAG) aims to generate more reliable and accurate responses, by augmenting large language models(LLMs) with the external vast and dynamic knowledge. Most previous work focuses on using RAG for single-round question answering, while how to adapt RAG to the complex conversational setting wherein the question is interdependent on the preceding context is not well studied. In this paper, we propose a conversation-level RAG (ConvRAG) approach, which incorporates fine-grained retrieval augmentation and self-check for conversational question answering (CQA). In particular, our approach consists of three components, namely conversational question refiner, fine-grained retriever and self-check based response generator, which work collaboratively for question understanding and relevant information acquisition in conversational settings. Extensive experiments demonstrate the great advantages of our approach over the state-of-the-art baselines. Moreover, we also release a Chinese CQA dataset with new features including reformulated question, extracted keyword, retrieved paragraphs and their helpfulness, which facilitates further researches in RAG enhanced CQA.

Graph partitioning is crucial in distributed graph-parallel computing systems, and it is challenging for graph partitioning to optimize the communication cost and load balancing together. Existing state-of-the-art works, such as Powerlyra and TopoX, optimize the load balancing by randomly distributing the edges of high-degree vertices, which inevitably brings a high communication cost that is unbounded. This paper proposes a graph partition model that can minimize communication cost while maximizing load balancing. More specifically, we model the graph partition as the combinatorial design problem. Our proposed model can provide high-quality partition that guarantees that the computing load can be evenly distributed to each worker and minimizes the communication cost with a near-optimal theoretical boundary.Based on the proposed model, we extend the hybrid-cut partitioning algorithm for the power-law graph and propose HCPD, a hybrid-cut partitioning algorithm based on combinatorial design. HCPD uses the proposed model to optimize the load balancing and communication cost simultaneously for high-degree vertices, and assigns the high-degree vertices and their low-degree neighbors to the same workers by label propagation to reduce the overall communication cost. In this way, we partition the low-degree and high-degree vertices holistically and further improve the partition quality, unlike Powerlyra and TopoX, which deal with the two parts independently. Our experiments show that HCPD outperforms Powerlyra on PageRank task by up to 2× faster on real-world power-law graphs with billions of edges.

Previous studies about event-level sentiment analysis (SA) usually model the event as a topic, a category or target terms, while the structured arguments (e.g., subject, object, time and location) that have potential effects on the sentiment are not well studied. In this paper, we redefine the task as structured event-level SA and propose an End-to-End Event-level Sentiment Analysis (E3SA) approach to solve this issue. Specifically, we explicitly extract and model the event structure information for enhancing event-level SA. Extensive experiments demonstrate the great advantages of our proposed approach over the state-of-the-art methods. Noting the lack of the dataset, we also release a large-scale real-world dataset with event arguments and sentiment labelling for promoting more researches.

We perform fast single image super-resolution with flexible magnification for natural images. A novel coarse-to-fine super-resolution framework is developed for the magnification that is factorized into a maximum integer component and the quotient. Specifically, our framework is embedded with a light-weight upscale network for super-resolution with the integer scale factor, followed by the fine-grained network to guide interpolation on feature maps as well as to generate the super-resolved image. Compared with the previous flexible magnification super-resolution approaches, the proposed framework achieves a tradeoff between computational complexity and performance. We conduct experiments using the coarse-to-fine framework on the standard benchmarks and demonstrate its superiority in terms of effectiveness and efficiency over previous approaches.

Knowledge graphs have been used in a wide range of applications to support search, recommendation, and question answering (Q&A). For example, in Q&A systems, given a new question, we may use a knowledge graph to automatically identify the most suitable answers based on similarity evaluation. However, such systems may suffer from two major limitations. First, the knowledge graph constructed based on source data may contain errors. Second, the knowledge graph may become out of date and cannot quickly adapt to new knowledge. To address these issues, in this paper, we propose an interactive framework that refines and optimizes knowledge graphs through user votes. We develop an efficient similarity evaluation notion, called extended inverse P-distance, based on which the graph optimization problem can be formulated as a signomial geometric programming problem. We then propose a basic single-vote solution and a more advanced multi-vote solution for graph optimization. We also propose a split-and-merge optimization strategy to scale up the multi-vote solution. Extensive experiments based on real-life and synthetic graphs demonstrate the effectiveness and efficiency of our proposed framework.

Although neural networks achieve promising performance in sentence level sentiment classification, most of them are not aware of sentiment commonsense, such as sentiment polarity tags (Positive or Negative) for words, which explicitly determine the sentiment of the sentence in most cases. In this paper, we propose an auxiliary tagging task to integrate sentiment commonsense into sequential neural networks (such as LSTM). We employ the advantage of multitask learning to achieve two goals simultaneously: 1) the sequential learning task accounts for incorporating the semantic information of the surrounding words; 2) the word tagging task ensures the sequential representation still retains the corresponding word tagging information. Besides, considering the most direct way to introduce sentiment information into models as additional knowledge, we further incorporate the additional knowledge enhancing tagging task model to strengthen the effect of sentiment commonsense. We prove the effectiveness of the sentiment commonsense by extensive experiments. The results show that our models exhibit consistent superiority over competitors on three real-word datasets. Specifically, we obtain an accuracy of 55.2%, which is a new state-of-the-art for SST-fine dataset.

The accuracy of OCR is usually affected by the quality of the input document image and different kinds of marred document images hamper the OCR results. Among these scenarios, the low-resolution image is a common and challenging case. In this paper, we propose the cascaded networks for document image super-resolution. Our model is composed by the Detail-Preserving Networks with small magnification. The loss function with perceptual terms is designed to simultaneously preserve the original patterns and enhance the edge of the characters. These networks are trained with the same architecture and different parameters and then assembled into a pipeline model with a larger magnification. The low-resolution images can upscale gradually by passing through each Detail-Preserving Network until the final high-resolution images. Through extensive experiments on two scanning document image datasets, we demonstrate that the proposed approach outperforms recent state-of-the-art image super-resolution methods, and combining it with standard OCR system lead to signification improvements on the recognition results.

Locating actions in long untrimmed videos has been a challenging problem in video content analysis. The performances of existing action localization approaches remain unsatisfactory in precisely determining the beginning and the end of an action. Imitating the human perception procedure with observations and refinements, we propose a novel three-phase action localization framework. Our framework is embedded with an Actionness Network to generate initial proposals through frame-wise similarity grouping, and then a Refinement Network to conduct boundary adjustment on these proposals. Finally, the refined proposals are sent to a Localization Network for further fine-grained location regression. The whole process can be deemed as multi-stage refinement using a novel non-local pyramid feature under various temporal granularities. We evaluate our framework on THUMOS14 benchmark and obtain a significant improvement over the state-of-the-arts approaches. Specifically, the performance gain is remarkable under precise localization with high IoU thresholds. Our proposed framework achieves [email protected]=0.5 of 34.2%.

The neural networks have attracted great attention for sentence similarity modeling in recent years. Most neural networks focus on the representation of each sentence, while the common features of a sentence pair are not well studied. In this paper, we propose a Collaborative and Adversarial Network (CAN), which explicitly models the common features between two sentences for enhancing sentence similarity modeling. To be specific, a common feature extractor is presented and embedded into our CAN model, which includes a generator and a discriminator playing a collaborative and adversarial game for common feature extraction. Experiments on three benchmark datasets, namely TREC-QA and WikiQA for answer selection and MSRP for paraphrase identification, show that our proposed model is effective to boost the performance of sentence similarity modeling. In particular, our proposed model outperforms the state-of-the-art approaches on TREC-QA without using any external resources or pre-training. For the other two datasets, our model is also comparable to if not better than the recent neural network approaches.

Query expansion under the pseudo-relevance feedback (PRF) framework has been extensively studied in information retrieval. However, most expansion methods are mainly based on the statistics of single terms, which can generate plenty of irrelevant query terms and decrease retrieval performance. To alleviate this problem, we propose an approach that adapts the PRF-based contextual snippets into a context-aware topic model to enhance query representations. Specifically, instead of selecting a series of independent terms, we make full use of the query contextual information and focus on the snippets with the length ofnin the PRF documents. Furthermore, we propose a context-aware topic (CAT) model to mine the topic distributions of the query-relevant snippets, namely, fine contextual snippets. In contrast to the traditional topic models that infer the topics from the whole corpus, we establish a bridge between the snippets and the corresponding PRF documents, which can be used for modeling the topics more precisely and efficiently. Finally, the topic distributions of the fine snippets are used for context-aware and topic-sensitive query representations. To evaluate the performance of our approach, we integrate the obtained queries into a topic-based hybrid retrieval model and conduct extensive experiments on various TREC collections. The experimental results show that our query-modeling approach is more effective in boosting retrieval performance compared with the state-of-the-art methods.

Temporal information has been widely used to promote the information retrieval (IR) performance, especially for microblog search which usually prefers the latest news and events. Previous studies mainly focused on incorporating the document-level temporal information into retrieval, while the temporal relevance of each query word was not well investigated. In this paper, we propose a word temporal predictor to characterize the word-level temporal relevance by fine-grained time-aware kernel density estimation over the feedback documents. In addition, we present a fine-grained time-aware framework to integrate the proposed word temporal predictor with the traditional document temporal predictor for retrieval. Finally, we incorporate the framework into two state-of-the-art retrieval models, namely language model (LM) and BM25. The experimental results on the TREC 2011-2014 Microblog collections, show that our proposed word temporal predictor is effective to boost the retrieval performance within both LM and BM25 frameworks. In particular, we achieve significant improvements over the strong baselines with optimized settings in most cases. Furthermore, our fine-grained time-aware models with word temporal predictor are comparable to if not better than the state-of-the-art temporal retrieval models.

Clinical Decision Systems utilize patient profiles to search for relevant medical support. Existing work on medical search has a primary topic on query expansion, which enriches queries by adding more useful terms. However, it performs well when queries are concise. In this paper, we aim to tackle verbose queries and propose a new automatic query reformulation method, which not only considers query expansion, but also includes query reduction. This improves the searching performance by refining the query to effectively avoid irrelevant results. The method is achieved by classifying each sentence into expansion and reduction categories with a weighted score model, which depends on the occurrence of medical and negative terms. The refined queries show promising results on experiments with TREC CDS datasets. In particular, the final performance makes improvements in terms of NDCG as 8.2% compared to the prevailing query expansion method, and 22.07% compared to the baseline of original query.

In this paper, we introduce TaiChi which is a fine-grained action dataset. It consists of unconstrained user-uploaded web videos containing camera motion and partial occlusions which pose new challenges to fine-grained action recognition compared to the existing datasets. In this dataset, 2,772 samples of 58 fine-grained action classes are manually annotated. Additionally, we provide the baseline action recognition results using the state-of-the-art Improved Dense Trajectory feature and Fisher Vector representation with an MAP (Mean Average Precision) of 51.39%.

Attention based recurrent neural networks (RNN) have shown a great success for question answering (QA) in recent years. Although significant improvements have been achieved over the non-attentive models, the position information is not well studied within the attention-based framework. Motivated by the effectiveness of using the word positional context to enhance information retrieval, we assume that if a word in the question (i.e., question word) occurs in an answer sentence, the neighboring words should be given more attention since they intuitively contain more valuable information for question answering than those far away. Based on this assumption, we propose a positional attention based RNN model, which incorporates the positional context of the question words into the answers' attentive representations. Experiments on two benchmark datasets show the great advantages of our proposed model. Specifically, we achieve a maximum improvement of 8.83% over the classical attention based RNN model in terms of mean average precision. Furthermore, our model is comparable to if not better than the state-of-the-art approaches for question answering.

Here we propose an advance Skip-gram model to incorporate both word sentiment and negation information. In particular, there is a a softmax layer for the word sentiment polarity upon the Skip-gram model. Then, two paralleled embedding layers are set up in the same embedding space, one for the affirmative context and the other for the negated context, followed by their loss functions. We evaluate our proposed model on the 2013 and 2014 SemEval data sets. The experimental results show that the proposed approach achieves better performance and learns higher dimensional word embedding informatively on the large-scale data.

Our slogan for the proposed Clairvoyant system is "with several clicks, the future is in your hand, the plan comes into your mind". Clairvoyant is to predict the future of new videos with only few data. The core function in the system is the novel shifted shape match prediction algorithm, based on a K-Nearest Neighbor model. Tons of experiments have been conducted on the open data sets. The experimental results confirms that the proposed SSMP algorithm is promising and outperforms the baselines with significant improvements on various evaluation methods. A demonstration video has been published at http://1drv.ms/1nyH3hD.