The Internet contains a wealth of knowledge -- from the birthdays of historical figures to tutorials on how to code -- all of which may be learned by language models. However, while certain pieces of information are ubiquitous on the web, others appear extremely rarely. In this paper, we study the relationship between the knowledge memorized by large language models and the information in pre-training datasets scraped from the web. In particular, we show that a language model's ability to answer a fact-based question relates to how many documents associated with that question were seen during pre-training. We identify these relevant documents by entity linking pre-training datasets and counting documents that contain the same entities as a given question-answer pair. Our results demonstrate strong correlational and causal relationships between accuracy and relevant document count for numerous question answering datasets (e.g., TriviaQA), pre-training corpora (e.g., ROOTS), and model sizes (e.g., 176B parameters). Moreover, while larger models are better at learning long-tail knowledge, we estimate that today's models must be scaled by many orders of magnitude to reach competitive QA performance on questions with little support in the pre-training data. Finally, we show that retrieval-augmentation can reduce the dependence on relevant pre-training information, presenting a promising approach for capturing the long-tail.
相關內容
Gesture recognition is a pivotal technology in the realm of intelligent education, and millimeter-wave (mmWave) signals possess advantages such as high resolution and strong penetration capability. This paper introduces a highly accurate and robust gesture recognition method using mmWave radar. The method involves capturing the raw signals of hand movements with the mmWave radar module and preprocessing the received radar signals, including Fourier transformation, distance compression, Doppler processing, and noise reduction through moving target indication (MTI). The preprocessed signals are then fed into the Convolutional Neural Network-Time Domain Convolutional Network (CNN-TCN) model to extract spatio-temporal features, with recognition performance evaluated through classification. Experimental results demonstrate that this method achieves an accuracy rate of 98.2% in domain-specific recognition and maintains a consistently high recognition rate across different neural networks, showcasing exceptional recognition performance and robustness.
Dysarthria is a speech disorder that hinders communication due to difficulties in articulating words. Detection of dysarthria is important for several reasons as it can be used to develop a treatment plan and help improve a person's quality of life and ability to communicate effectively. Much of the literature focused on improving ASR systems for dysarthric speech. The objective of the current work is to develop models that can accurately classify the presence of dysarthria and also give information about the intelligibility level using limited data by employing a few-shot approach using a transformer model. This work also aims to tackle the data leakage that is present in previous studies. Our whisper-large-v2 transformer model trained on a subset of the UASpeech dataset containing medium intelligibility level patients achieved an accuracy of 85%, precision of 0.92, recall of 0.8 F1-score of 0.85, and specificity of 0.91. Experimental results also demonstrate that the model trained using the 'words' dataset performed better compared to the model trained on the 'letters' and 'digits' dataset. Moreover, the multiclass model achieved an accuracy of 67%.
The detection of spoofing speech generated by unseen algorithms remains an unresolved challenge. One reason for the lack of generalization ability is traditional detecting systems follow the binary classification paradigm, which inherently assumes the possession of prior knowledge of spoofing speech. One-class methods attempt to learn the distribution of bonafide speech and are inherently suited to the task where spoofing speech exhibits significant differences. However, training a one-class system using only bonafide speech is challenging. In this paper, we introduce a teacher-student framework to provide guidance for the training of a one-class model. The proposed one-class knowledge distillation method outperforms other state-of-the-art methods on the ASVspoof 21DF dataset and InTheWild dataset, which demonstrates its superior generalization ability.
Smishing, also known as SMS phishing, is a type of fraudulent communication in which an attacker disguises SMS communications to deceive a target into providing their sensitive data. Smishing attacks use a variety of tactics; however, they have a similar goal of stealing money or personally identifying information (PII) from a victim. In response to these attacks, a wide variety of anti-smishing tools have been developed to block or filter these communications. Despite this, the number of phishing attacks continue to rise. In this paper, we developed a test bed for measuring the effectiveness of popular anti-smishing tools against fresh smishing attacks. To collect fresh smishing data, we introduce Smishtank.com, a collaborative online resource for reporting and collecting smishing data sets. The SMS messages were validated by a security expert and an in-depth qualitative analysis was performed on the collected messages to provide further insights. To compare tool effectiveness, we experimented with 20 smishing and benign messages across 3 key segments of the SMS messaging delivery ecosystem. Our results revealed significant room for improvement in all 3 areas against our smishing set. Most anti-phishing apps and bulk messaging services didn't filter smishing messages beyond the carrier blocking. The 2 apps that blocked the most smish also blocked 85-100\% of benign messages. Finally, while carriers did not block any benign messages, they were only able to reach a 25-35\% blocking rate for smishing messages. Our work provides insights into the performance of anti-smishing tools and the roles they play in the message blocking process. This paper would enable the research community and industry to be better informed on the current state of anti-smishing technology on the SMS platform.
Despite their impressive performance in a wide range of NLP tasks, Large Language Models (LLMs) have been reported to encode worrying-levels of gender bias. Prior work has proposed debiasing methods that require human labelled examples, data augmentation and fine-tuning of the LLMs, which are computationally costly. Moreover, one might not even have access to the internal parameters for performing debiasing such as in the case of commercially available LLMs such as GPT-4. To address this challenge we propose bias suppression, a novel alternative to debiasing that does not require access to model parameters. We show that text-based preambles, generated from manually designed templates covering counterfactual statements, can accurately suppress gender biases in LLMs. Moreover, we find that descriptive sentences for occupations can further suppress gender biases. Interestingly, we find that bias suppression has a minimal adverse effect on downstream task performance, while effectively mitigating the gender biases.
Emotion recognition in conversation (ERC) aims to detect the emotion label for each utterance. Motivated by recent studies which have proven that feeding training examples in a meaningful order rather than considering them randomly can boost the performance of models, we propose an ERC-oriented hybrid curriculum learning framework. Our framework consists of two curricula: (1) conversation-level curriculum (CC); and (2) utterance-level curriculum (UC). In CC, we construct a difficulty measurer based on "emotion shift" frequency within a conversation, then the conversations are scheduled in an "easy to hard" schema according to the difficulty score returned by the difficulty measurer. For UC, it is implemented from an emotion-similarity perspective, which progressively strengthens the model's ability in identifying the confusing emotions. With the proposed model-agnostic hybrid curriculum learning strategy, we observe significant performance boosts over a wide range of existing ERC models and we are able to achieve new state-of-the-art results on four public ERC datasets.
Recent VQA models may tend to rely on language bias as a shortcut and thus fail to sufficiently learn the multi-modal knowledge from both vision and language. In this paper, we investigate how to capture and mitigate language bias in VQA. Motivated by causal effects, we proposed a novel counterfactual inference framework, which enables us to capture the language bias as the direct causal effect of questions on answers and reduce the language bias by subtracting the direct language effect from the total causal effect. Experiments demonstrate that our proposed counterfactual inference framework 1) is general to various VQA backbones and fusion strategies, 2) achieves competitive performance on the language-bias sensitive VQA-CP dataset while performs robustly on the balanced VQA v2 dataset.
Few-shot Knowledge Graph (KG) completion is a focus of current research, where each task aims at querying unseen facts of a relation given its few-shot reference entity pairs. Recent attempts solve this problem by learning static representations of entities and references, ignoring their dynamic properties, i.e., entities may exhibit diverse roles within task relations, and references may make different contributions to queries. This work proposes an adaptive attentional network for few-shot KG completion by learning adaptive entity and reference representations. Specifically, entities are modeled by an adaptive neighbor encoder to discern their task-oriented roles, while references are modeled by an adaptive query-aware aggregator to differentiate their contributions. Through the attention mechanism, both entities and references can capture their fine-grained semantic meanings, and thus render more expressive representations. This will be more predictive for knowledge acquisition in the few-shot scenario. Evaluation in link prediction on two public datasets shows that our approach achieves new state-of-the-art results with different few-shot sizes.
Explainable Recommendation refers to the personalized recommendation algorithms that address the problem of why -- they not only provide the user with the recommendations, but also make the user aware why such items are recommended by generating recommendation explanations, which help to improve the effectiveness, efficiency, persuasiveness, and user satisfaction of recommender systems. In recent years, a large number of explainable recommendation approaches -- especially model-based explainable recommendation algorithms -- have been proposed and adopted in real-world systems. In this survey, we review the work on explainable recommendation that has been published in or before the year of 2018. We first high-light the position of explainable recommendation in recommender system research by categorizing recommendation problems into the 5W, i.e., what, when, who, where, and why. We then conduct a comprehensive survey of explainable recommendation itself in terms of three aspects: 1) We provide a chronological research line of explanations in recommender systems, including the user study approaches in the early years, as well as the more recent model-based approaches. 2) We provide a taxonomy for explainable recommendation algorithms, including user-based, item-based, model-based, and post-model explanations. 3) We summarize the application of explainable recommendation in different recommendation tasks, including product recommendation, social recommendation, POI recommendation, etc. We devote a chapter to discuss the explanation perspectives in the broader IR and machine learning settings, as well as their relationship with explainable recommendation research. We end the survey by discussing potential future research directions to promote the explainable recommendation research area.
We investigate the problem of automatically determining what type of shoe left an impression found at a crime scene. This recognition problem is made difficult by the variability in types of crime scene evidence (ranging from traces of dust or oil on hard surfaces to impressions made in soil) and the lack of comprehensive databases of shoe outsole tread patterns. We find that mid-level features extracted by pre-trained convolutional neural nets are surprisingly effective descriptors for this specialized domains. However, the choice of similarity measure for matching exemplars to a query image is essential to good performance. For matching multi-channel deep features, we propose the use of multi-channel normalized cross-correlation and analyze its effectiveness. Our proposed metric significantly improves performance in matching crime scene shoeprints to laboratory test impressions. We also show its effectiveness in other cross-domain image retrieval problems: matching facade images to segmentation labels and aerial photos to map images. Finally, we introduce a discriminatively trained variant and fine-tune our system through our proposed metric, obtaining state-of-the-art performance.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191