Video question answering involves understanding video content to generate accurate answers to questions. Recent studies have successfully modeled video features and achieved diverse multimodal interaction, yielding impressive outcomes. However, they have overlooked the fact that the video contains richer instances and events beyond the scope of the stated question. Extremely imbalanced alignment of information from both sides leads to significant instability in reasoning. To address this concern, we propose the Video-Context Aligned Transformer (V-CAT), which leverages the context to achieve semantic and content alignment between video and question. Specifically, the video and text are encoded into a shared semantic space initially. We apply contrastive learning to global video token and context token to enhance the semantic alignment. Then, the pooled context feature is utilized to obtain corresponding visual content. Finally, the answer is decoded by integrating the refined video and question features. We evaluate the effectiveness of V-CAT on MSVD-QA and MSRVTT-QA dataset, both achieving state-of-the-art performance. Extended experiments further analyze and demonstrate the effectiveness of each proposed module.
论文检索
输入标题、作者或关键词，从 2,866 篇学术成果中精准定位
Towards Explainable Joint Models via Information Theory for Multiple Intent Detection and Slot Filling
PDF ↗Recent joint models for multi-intent detection and slot filling have obtained promising results through modeling the unidirectional or bidirectional guidance between intent and slot. However, existing works design joint models heuristically and lack some theoretical exploration, including (1) theoretical measurement of the joint-interaction quality; (2) explainability of design and optimization methods of joint models, which may limit the performance and efficiency of designs. In this paper, we mathematically define the cross-task information gain (CIG) to measure the quality of joint processes from an information-theoretic perspective and discover an implicit optimization of CIG in previous models. Based on this, we propose a novel multi-stage iterative framework with theoretical effectiveness, explainability, and convergence, which can explicitly optimize information for cross-task interactions. Further, we devise an information-based joint model (InfoJoint) that conforms to this theoretical framework to gradually reduce the cross-task propagation of erroneous semantics through CIG iterative maximization. Extensive experiment results on two public datasets show that InfoJoint outperforms the state-of-the-art models by a large margin.
Aligner²: Enhancing Joint Multiple Intent Detection and Slot Filling via Adjustive and Forced Cross-Task Alignment
PDF ↗Multi-intent spoken language understanding (SLU) has garnered growing attention due to its ability to handle multiple intent utterances, which closely mirrors practical scenarios. Unlike traditional SLU, each intent in multi-intent SLU corresponds to its designated scope for slots, which occurs in certain fragments within the utterance. As a result, establishing precise scope alignment to mitigate noise impact emerges as a key challenge in multi-intent SLU. More seriously, they lack alignment between the predictions of the two sub-tasks due to task-independent decoding, resulting in a limitation on the overall performance. To address these challenges, we propose a novel framework termed Aligner² for multi-intent SLU, which contains an Adjustive Cross-task Aligner (ACA) and a Forced Cross-task Aligner (FCA). ACA utilizes the information conveyed by joint label embeddings to accurately align the scope of intent and corresponding slots, before the interaction of the two subtasks. FCA introduces reinforcement learning, to enforce the alignment of the task-specific hidden states after the interaction, which is explicitly guided by the prediction. Extensive experiments on two public multi-intent SLU datasets demonstrate the superiority of our Aligner² over state-of-the-art methods. More encouragingly, the proposed method Aligner² can be easily integrated into existing multi-intent SLU frameworks, to further boost performance.
Multichannel AV-wav2vec2: A Framework for Learning Multichannel Multi-Modal Speech Representation
PDF ↗Self-supervised speech pre-training methods have developed rapidly in recent years, which show to be very effective for many near-field single-channel speech tasks. However, far-field multichannel speech processing is suffering from the scarcity of labeled multichannel data and complex ambient noises. The efficacy of self-supervised learning for far-field multichannel and multi-modal speech processing has not been well explored. Considering that visual information helps to improve speech recognition performance in noisy scenes, in this work we propose the multichannel multi-modal speech self-supervised learning framework AV-wav2vec2, which utilizes video and multichannel audio data as inputs. First, we propose a multi-path structure to process multi-channel audio streams and a visual stream in parallel, with intra-, and inter-channel contrastive as training targets to fully exploit the rich information in multi-channel speech data. Second, based on contrastive learning, we use additional single-channel audio data, which is trained jointly to improve the performance of multichannel multi-modal representation. Finally, we use a Chinese multichannel multi-modal dataset in real scenarios to validate the effectiveness of the proposed method on audio-visual speech recognition (AVSR), automatic speech recognition (ASR), visual speech recognition (VSR) and audio-visual speaker diarization (AVSD) tasks.
Natural language processing models are vulnerable to adversarial examples. Previous textual adversarial attacks adopt model internal information (gradients or confidence scores) to generate adversarial examples. However, this information is unavailable in the real world. Therefore, we focus on a more realistic and challenging setting, named hard-label attack, in which the attacker can only query the model and obtain a discrete prediction label. Existing hard-label attack algorithms tend to initialize adversarial examples by random substitution and then utilize complex heuristic algorithms to optimize the adversarial perturbation. These methods require a lot of model queries and the attack success rate is restricted by adversary initialization. In this paper, we propose a novel hard-label attack algorithm named LimeAttack, which leverages a local explainable method to approximate word importance ranking, and then adopts beam search to find the optimal solution. Extensive experiments show that LimeAttack achieves the better attacking performance compared with existing hard-label attack under the same query budget. In addition, we evaluate the effectiveness of LimeAttack on large language models and some defense methods, and results indicate that adversarial examples remain a significant threat to large language models. The adversarial examples crafted by LimeAttack are highly transferable and effectively improve model robustness in adversarial training.
With the boom of Large Language Models (LLMs), the research of solving Math Word Problem (MWP) has recently made great progress. However, there are few studies to examine the robustness of LLMs in math solving ability. Instead of attacking prompts in the use of LLMs, we propose a MathAttack model to attack MWP samples which are closer to the essence of robustness in solving math problems. Compared to traditional text adversarial attack, it is essential to preserve the mathematical logic of original MWPs during the attacking. To this end, we propose logical entity recognition to identify logical entries which are then frozen. Subsequently, the remaining text are attacked by adopting a word-level attacker. Furthermore, we propose a new dataset RobustMath to evaluate the robustness of LLMs in math solving ability. Extensive experiments on our RobustMath and two another math benchmark datasets GSM8K and MultiAirth show that MathAttack could effectively attack the math solving ability of LLMs. In the experiments, we observe that (1) Our adversarial samples from higher-accuracy LLMs are also effective for attacking LLMs with lower accuracy (e.g., transfer from larger to smaller-size LLMs, or from few-shot to zero-shot prompts); (2) Complex MWPs (such as more solving steps, longer text, more numbers) are more vulnerable to attack; (3) We can improve the robustness of LLMs by using our adversarial samples in few-shot prompts. Finally, we hope our practice and observation can serve as an important attempt towards enhancing the robustness of LLMs in math solving ability. The code and dataset is available at: https://github.com/zhouzihao501/MathAttack.
Large language models (LLMs) have been proven capable of memorizing their training data, which can be extracted through specifically designed prompts. As the scale of datasets continues to grow, privacy risks arising from memorization have attracted increasing attention. Quantifying language model memorization helps evaluate potential privacy risks. However, prior works on quantifying memorization require access to the precise original data or incur substantial computational overhead, making it difficult for applications in real-world language models. To this end, we propose a fine-grained, entity-level definition to quantify memorization with conditions and metrics closer to real-world scenarios. In addition, we also present an approach for efficiently extracting sensitive entities from autoregressive language models. We conduct extensive experiments based on the proposed, probing language models' ability to reconstruct sensitive entities under different settings. We find that language models have strong memorization at the entity level and are able to reproduce the training data even with partial leakages. The results demonstrate that LLMs not only memorize their training data but also understand associations between entities. These findings necessitate that trainers of LLMs exercise greater prudence regarding model memorization, adopting memorization mitigation techniques to preclude privacy violations.
Long document retrieval aims to fetch query-relevant documents from a large-scale collection, where knowledge distillation has become de facto to improve a retriever by mimicking a heterogeneous yet powerful cross-encoder. However, in contrast to passages or sentences, retrieval on long documents suffers from the \textit{scope hypothesis} that a long document may cover multiple topics. This maximizes their structure heterogeneity and poses a granular-mismatch issue, leading to an inferior distillation efficacy. In this work, we propose a new learning framework, fine-grained distillation (FGD), for long-document retrievers. While preserving the conventional dense retrieval paradigm, it first produces global-consistent representations crossing different fine granularity and then applies multi-granular aligned distillation merely during training. In experiments, we evaluate our framework on two long-document retrieval benchmarks, which show state-of-the-art performance.
Large Language Models (LLMs) have drastically reshaped our interactions with artificial intelligence (AI) systems, showcasing impressive performance across an extensive array of tasks. Despite this, a notable hindrance remains—the deficiency of a long-term memory mechanism within these models. This shortfall becomes increasingly evident in situations demanding sustained interaction, such as personal companion systems, psychological counseling, and secretarial assistance. Recognizing the necessity for long-term memory, we propose MemoryBank, a novel memory mechanism tailored for LLMs. MemoryBank enables the models to summon relevant memories, continually evolve through continuous memory updates, comprehend, and adapt to a user's personality over time by synthesizing information from previous interactions. To mimic anthropomorphic behaviors and selectively preserve memory, MemoryBank incorporates a memory updating mechanism, inspired by the Ebbinghaus Forgetting Curve theory. This mechanism permits the AI to forget and reinforce memory based on time elapsed and the relative significance of the memory, thereby offering a more human-like memory mechanism and enriched user experience. MemoryBank is versatile in accommodating both closed-source models like ChatGPT and open-source models such as ChatGLM. To validate MemoryBank's effectiveness, we exemplify its application through the creation of an LLM-based chatbot named SiliconFriend in a long-term AI Companion scenario. Further tuned with psychological dialog data, SiliconFriend displays heightened empathy and discernment in its interactions. Experiment involves both qualitative analysis with real-world user dialogs and quantitative analysis with simulated dialogs. In the latter, ChatGPT acts as multiple users with diverse characteristics and generates long-term dialog contexts covering a wide array of topics. The results of our analysis reveal that SiliconFriend, equipped with MemoryBank, exhibits a strong capability for long-term companionship as it can provide emphatic response, recall relevant memories and understand user personality.
You Only Read Once: Constituency-Oriented Relational Graph Convolutional Network for Multi-Aspect Multi-Sentiment Classification
PDF ↗Most of the existing aspect-based sentiment analysis (ABSA) models only predict the sentiment polarity of a single aspect at a time, focusing primarily on enhancing the representation of this single aspect based on the other contexts or aspects. This one-to-one paradigm ignores the fact that multi-aspect, multi-sentiment sentences contain not only distinct specific descriptions for distinct specific aspects, but also shared global context information for multiple aspects. To fully consider these issues, we propose a one-to-many ABSA framework, called You Only Read Once (YORO), that can simultaneously model representations of all aspects based on their specific descriptions and better fuse their relationships using globally shared contextual information in the sentence. Predicting the sentiment polarity of multiple aspects simultaneously is beneficial to improving the efficacy of calculation and prediction. Extensive experiments are conducted on three public datasets (MAMS, Rest14, and Lap14). Experimental results demonstrate the effectiveness of YORO in handling multi-aspect, multi-sentiment scenarios and highlight the promise of one-to-many ABSA in balancing efficiency and accuracy.
Existing neural models are demonstrated to struggle with compositional generalization (CG), i.e., the ability to systematically generalize to unseen compositions of seen components. A key reason for failure on CG is that the syntactic and semantic representations of sequences in both the uppermost layer of the encoder and decoder are entangled. However, previous work concentrates on separating the learning of syntax and semantics instead of exploring the reasons behind the representation entanglement (RE) problem to solve it. We explain why it exists by analyzing the representation evolving mechanism from the bottom to the top of the Transformer layers. We find that the ``shallow'' residual connections within each layer fail to fuse previous layers' information effectively, leading to information forgetting between layers and further the RE problems. Inspired by this, we propose LRF, a novel Layer-wise Representation Fusion framework for CG, which learns to fuse previous layers' information back into the encoding and decoding process effectively through introducing a fuse-attention module at each encoder and decoder layer. LRF achieves promising results on two realistic benchmarks, empirically demonstrating the effectiveness of our proposal. Codes are available at https://github.com/thinkaboutzero/LRF.
Although singing voice synthesis (SVS) has made significant progress recently, with its unique styles and various genres, Chinese opera synthesis requires greater attention but is rarely studied for lack of training data and high expressiveness. In this work, we build a high-quality Gezi Opera (a type of Chinese opera popular in Fujian and Taiwan) audio-text alignment dataset and formulate specific data annotation methods applicable to Chinese operas. We propose FT-GAN, an acoustic model for fine-grained tune modeling in Chinese opera synthesis based on the empirical analysis of the differences between Chinese operas and pop songs. To further improve the quality of the synthesized opera, we propose a speech pre-training strategy for additional knowledge injection. The experimental results show that FT-GAN outperforms the strong baselines in SVS on the Gezi Opera synthesis task. Extensive experiments further verify that FT-GAN performs well on synthesis tasks of other operas such as Peking Opera. Audio samples, the dataset, and the codes are available at https://zhengmidon.github.io/FTGAN.github.io/.
With the proliferation of dialogic data across the Internet, the Dialogue Commonsense Multi-choice Question Answering (DC-MCQ) task has emerged as a response to the challenge of comprehending user queries and intentions. Although prevailing methodologies exhibit effectiveness in addressing single-choice questions, they encounter difficulties in handling multi-choice queries due to the heightened intricacy and informational density. In this paper, inspired by the human cognitive process of progressively excluding options, we propose a three-step Reverse Exclusion Graph-of-Thought (ReX-GoT) framework, including Option Exclusion, Error Analysis, and Combine Information. Specifically, our ReX-GoT mimics human reasoning by gradually excluding irrelevant options and learning the reasons for option errors to choose the optimal path of the GoT and ultimately infer the correct answer. By progressively integrating intricate clues, our method effectively reduces the difficulty of multi-choice reasoning and provides a novel solution for DC-MCQ. Extensive experiments on the CICERO and CICERO_v2 datasets validate the significant improvement of our approach on DC-MCQ task. On zero-shot setting, our model outperform the best baseline by 17.67% in terms of F1 score for the multi-choice task. Most strikingly, our GPT3.5-based ReX-GoT framework achieves a remarkable 39.44% increase in F1 score.
Recently, lots of works that incorporate external lexicon information into character-level Chinese named entity recognition(NER) to overcome the lackness of natural delimiters of words, have achieved many advanced performance. However, obtaining and maintaining high-quality lexicons is costly, especially in special domains. In addition, the entity boundary bias caused by high mention coverage in some boundary characters poses a significant challenge to the generalization of NER models but receives little attention in the existing literature. To address these issues, we propose SENCR, a Span Enhanced Two-stage Network with Counterfactual Rethinking for Chinese NER, that contains a boundary detector for boundary supervision, a convolution-based type classifier for better span representation and a counterfactual rethinking(CR) strategy for debiased boundary detection in inference. The proposed boundary detector and type classifier are jointly trained with the same contextual encoder and then the trained boundary detector is debiased by our proposed CR strategy without modifying any model parameters in the inference stage. Extensive experiments on four Chinese NER datasets show the effectiveness of our proposed approach.
QPEN: Quantum Projection and Quantum Entanglement Enhanced Network for Cross-Lingual Aspect-Based Sentiment Analysis
PDF ↗Aspect-based sentiment analysis (ABSA) has attracted much attention due to its wide application scenarios. Most previous studies have focused solely on monolingual ABSA, posing a formidable challenge when extending ABSA applications to multilingual scenarios. In this paper, we study upgrading monolingual ABSA to cross-lingual ABSA. Existing methods usually exploit pre-trained cross-lingual language to model cross-lingual ABSA, and enhance the model with translation data. However, the low-resource languages might be under-represented during the pre-training phase, and the translation-enhanced methods heavily rely on the quality of the translation and label projection. Inspired by the observation that quantum entanglement can correlate multiple single systems, we map the monolingual expression to the quantum Hilbert space as a single quantum system, and then utilize quantum entanglement and quantum measurement to achieve cross-lingual ABSA. Specifically, we propose a novel quantum neural model named QPEN (short for quantum projection and quantum entanglement enhanced network). It is equipped with a proposed quantum projection module that projects aspects as quantum superposition on a complex-valued Hilbert space. Furthermore, a quantum entanglement module is proposed in QPEN to share language-specific features between different languages without transmission. We conducted simulation experiments on the classical computer, and experimental results on SemEval-2016 dataset demonstrate that our method achieves state-of-the-art performance in terms of F1-scores for five languages.
MultiSum: A Multi-Facet Approach for Extractive Social Summarization Utilizing Semantic and Sociological Relationships
PDF ↗Social summarization aims to provide summaries for a large number of social texts (called posts) about a single topic. To extract a summary, both the representation of post and summary selection method are crucial. Previous methods introduce social relation to enhance post embedding to mitigate the sparse representation due to its brief and informal expression. However, they ignore that there are multiple relations between posts. Besides, existing graph-based centrality calculation approaches tend to select posts from one aspect. This leads to facet bias especially when there are multiple viewpoints. In this paper, we propose a model named MultiSum to improve social summarization. Specifically, 1) We use graph convolutional networks to fuse text content with social and semantic relations to improve post representation; 2) The similarity between the summary and all aspects is incorporated into the centrality score during the selection phase, encouraging the model to pay attention to different facets. Experimental results on English and Chinese corpora support the effectiveness of this model. Furthermore, external evaluations by human experts and large language models demonstrate the validity of MultiSum in facet coverage and redundancy reduction.
Augmenting Language Models (LMs) with structured knowledge graphs (KGs) aims to leverage structured world knowledge to enhance the capability of LMs to complete knowledge-intensive tasks. However, existing methods are unable to effectively utilize the structured knowledge in a KG due to their inability to capture the rich relational semantics of knowledge triplets. Moreover, the modality gap between natural language text and KGs has become a challenging obstacle when aligning and fusing cross-modal information. To address these challenges, we propose a novel knowledge-augmented question answering (QA) model, namely, Graph Reasoning Transformers (GRT). Different from conventional node-level methods, the GRT serves knowledge triplets as atomic knowledge and utilize a triplet-level graph encoder to capture triplet-level graph features. Furthermore, to alleviate the negative effect of the modality gap on joint reasoning, we propose a representation alignment pretraining to align the cross-modal representations and introduce a cross-modal information fusion module with attention bias to enable fine-grained information fusion. Extensive experiments conducted on three knowledge-intensive QA benchmarks show that the GRT outperforms the state-of-the-art KG-augmented QA systems, demonstrating the effectiveness and adaptation of our proposed model.
Sign language translation (SLT) aims to convert continuous sign language videos into textual sentences. As a typical multi-modal task, there exists an inherent modality gap between sign language videos and spoken language text, which makes the cross-modal alignment between visual and textual modalities crucial. However, previous studies tend to rely on an intermediate sign gloss representation to help alleviate the cross-modal problem thereby neglecting the alignment across modalities that may lead to compromised results. To address this issue, we propose a novel framework based on Conditional Variational autoencoder for SLT (CV-SLT) that facilitates direct and sufficient cross-modal alignment between sign language videos and spoken language text. Specifically, our CV-SLT consists of two paths with two Kullback-Leibler (KL) divergences to regularize the outputs of the encoder and decoder, respectively. In the prior path, the model solely relies on visual information to predict the target text; whereas in the posterior path, it simultaneously encodes visual information and textual knowledge to reconstruct the target text. The first KL divergence optimizes the conditional variational autoencoder and regularizes the encoder outputs, while the second KL divergence performs a self-distillation from the posterior path to the prior path, ensuring the consistency of decoder outputs.We further enhance the integration of textual information to the posterior path by employing a shared Attention Residual Gaussian Distribution (ARGD), which considers the textual information in the posterior path as a residual component relative to the prior path. Extensive experiments conducted on public datasets demonstrate the effectiveness of our framework, achieving new state-of-the-art results while significantly alleviating the cross-modal representation discrepancy. The code and models are available at https://github.com/rzhao-zhsq/CV-SLT.
The recent surge in research interest in applying large language models (LLMs) to decision-making tasks has flourished by leveraging the extensive world knowledge embedded in LLMs. While there is a growing demand to tailor LLMs for custom decision-making tasks, finetuning them for specific tasks is resource-intensive and may diminish the model's generalization capabilities. Moreover, state-of-the-art language models like GPT-4 and Claude are primarily accessible through API calls, with their parametric weights remaining proprietary and unavailable to the public. This scenario emphasizes the growing need for new methodologies that allow learning from agent experiences without requiring parametric updates. To address these problems, we introduce the Experiential Learning (ExpeL) agent. Our agent autonomously gathers experiences and extracts knowledge using natural language from a collection of training tasks. At inference, the agent recalls its extracted insights and past experiences to make informed decisions. Our empirical results highlight the robust learning efficacy of the ExpeL agent, indicating a consistent enhancement in its performance as it accumulates experiences. We further explore the emerging capabilities and transfer learning potential of the ExpeL agent through qualitative observations and additional experiments.
Mind-map generation aims to process a document into a hierarchical structure to show its central idea and branches. Such a manner is more conducive to understanding the logic and semantics of the document than plain text. Recently, a state-of-the-art method encodes the sentences of a document sequentially and converts them to a relation graph via sequence-to-graph. Though this method is efficient to generate mind-maps in parallel, its mechanism focuses more on sequential features while hardly capturing structural information. Moreover, it's difficult to model long-range semantic relations. In this work, we propose a coreference-guided mind-map generation network (CMGN) to incorporate external structure knowledge. Specifically, we construct a coreference graph based on the coreference semantic relationship to introduce the graph structure information. Then we employ a coreference graph encoder to mine the potential governing relations between sentences. In order to exclude noise and better utilize the information of the coreference graph, we adopt a graph enhancement module in a contrastive learning manner. Experimental results demonstrate that our model outperforms all the existing methods. The case study further proves that our model can more accurately and concisely reveal the structure and semantics of a document. Code and data are available at https://github.com/Cyno2232/CMGN.