Existing scientific theories do not have the capability to operate on perception-based information. In a significant departure from existing methods, the high expressive power of naturallanguages is harnessed by cons...
详细信息
Existing scientific theories do not have the capability to operate on perception-based information. In a significant departure from existing methods, the high expressive power of naturallanguages is harnessed by constructing a precisiated naturallanguage (PNL). The PNL concept and the associated methodologies of computing with words and the computational theory of perceptions open the door to a wide-ranging generalization and restructuring of existing theories.
Multilingual language models (LMs) have become a powerful tool in NLP, especially for non-English languages. Nevertheless, model parameters of multilingual LMs remain large due to the larger embedding matrix of the vo...
详细信息
ISBN:
(纸本)9798891760615
Multilingual language models (LMs) have become a powerful tool in NLP, especially for non-English languages. Nevertheless, model parameters of multilingual LMs remain large due to the larger embedding matrix of the vocabulary covering tokens in different languages. Instead, monolingual LMs can be trained in a target language with the language-specific vocabulary only. In this paper, we propose vocabulary-trimming (VT), a method to reduce a multilingual LM vocabulary to a target language by deleting potentially irrelevant tokens from its vocabulary. In theory, VT can compress any existing multilingual LM to any language covered by the original model. In our experiments, we show that VT can retain the original performance of the multilingual LM, while being considerably smaller in size than the original multilingual LM. The evaluation is performed over four NLP tasks (two generative and two classification tasks) among four widely used multilingual LMs in seven languages. The results show that this methodology can keep the best of both monolingual and multilingual worlds by keeping a small size as monolingual models without the need for specifically retraining them, and can even help limit potentially harmful social biases.
As large language models are integrated into society, robustness toward a suite of prompts is increasingly important to maintain reliability in a high-variance environment. Robustness evaluations must comprehensively ...
详细信息
ISBN:
(纸本)9798891760615
As large language models are integrated into society, robustness toward a suite of prompts is increasingly important to maintain reliability in a high-variance environment. Robustness evaluations must comprehensively encapsulate the various settings in which a user may invoke an intelligent system. This paper proposes ASSERT, Automated Safety ScEnario Red Teaming, consisting of three methods - semantically aligned augmentation, target bootstrapping, and adversarial knowledge injection. For robust safety evaluation, we apply these methods in the critical domain of AI safety to algorithmically generate a test suite of prompts covering diverse robustness settings - semantic equivalence, related scenarios, and adversarial. We partition our prompts into four safety domains for a fine-grained analysis of how the domain affects model performance. Despite dedicated safeguards in existing state-of-the-art models, we find statistically significant performance differences of up to 11% in absolute classification accuracy among semantically related scenarios and error rates of up to 19% absolute error in zero-shot adversarial settings, raising concerns for users' physical safety.
Evaluation of NLP methods requires testing against a previously vetted gold-standard test set and reporting standard metrics (accuracy/precision/recall/F1). The current assumption is that all items in a given test set...
详细信息
We propose a Transformer-based sequence-to-sequence model for automatic speech recognition (ASR) capable of simultaneously transcribing and annotating audio with linguistic information such as phonemic transcripts or ...
详细信息
ISBN:
(纸本)9781954085466
We propose a Transformer-based sequence-to-sequence model for automatic speech recognition (ASR) capable of simultaneously transcribing and annotating audio with linguistic information such as phonemic transcripts or part-of-speech (POS) tags. Since linguistic information is important in naturallanguageprocessing (NLP), the proposed ASR is especially useful for speech interface applications, including spoken dialogue systems and speech translation, which combine ASR and NLP. To produce linguistic annotations, we train the ASR system using modified training targets: each grapheme or multi-grapheme unit in the target transcript is followed by an aligned phoneme sequence and/or POS tag. Since our method has access to the underlying audio data, we can estimate linguistic annotations more accurately than pipeline approaches in which NLP-based methods are applied to a hypothesized ASR transcript. Experimental results on Japanese and English datasets show that the proposed ASR system is capable of simultaneously producing high-quality transcriptions and linguistic annotations.
language understanding (LU) modules for spoken dialogue systems in the early phases of their development need to be (i) easy to construct and (ii) robust against various expressions. Conventional methods of LU are not...
详细信息
Software engineering is an intrinsically collaborative activity, especially in the era of Agile Software Development. Many actors are partaking in development activities, such that a common understanding should be rea...
详细信息
ISBN:
(纸本)9781728118765
Software engineering is an intrinsically collaborative activity, especially in the era of Agile Software Development. Many actors are partaking in development activities, such that a common understanding should be reached at numerous stages during the overall development life-cycle. For a few years now, naturallanguageprocessing techniques have been employed either to extract key information from free-form text or to generate models from the analysis of text in order to ease the sharing of knowledge across all parties. A significant part of these approaches focuses on retrieving lost domain and architectural knowledge through the analysis of documents, issue management systems or other forms of knowledge management systems. However, these post-processingmethods are time-consuming by nature since they require to invest significant resources into the validation of the extracted knowledge. In this paper, inspired by collaborative tools, bots and naturallanguage extraction approaches, we envision new ways to collaboratively record and document design decisions as they are discussed. These decisions will be documented as they are taken and, for some of them, static or behavioural models may be generated on-the-fly. Such an interactive process will ensure everyone agrees on critical design aspects of the software. We believe development teams will benefit from this approach because manual encoding of design knowledge will be reduced and will not be pushed to a later stage, when not forgotten.
Scene Text Recognition (STR) has long been considered an important yet challenging task in the field of computer vision. Recent works have demonstrated that utilizing language information is effective for the visually...
详细信息
ISBN:
(纸本)9798350344868;9798350344851
Scene Text Recognition (STR) has long been considered an important yet challenging task in the field of computer vision. Recent works have demonstrated that utilizing language information is effective for the visually difficult images, like ones with occultation or blurring. However, the use of language information sometimes leads to the over-correction problem. For out-of-vocabulary samples (e.g. "hou" and "0x4a"), some methods have tended to be biased to language side and over-corrected (e.g. over-correct "hou" to "hot"). This imbalance of vision and language has limited the usage of models in practical scenarios, yet it is rarely occurs for human. To address this issue, we rethink the human's recognition process and propose a model behaving in the order of "Read, Spell and Repeat". It refines the recognition process circularly with vision and language information. With this mechanism, our model integrates vision and language information in a more effective manner, achieving higher accuracy with less parameters compared to baseline and competitive performance with SOTA methods in the standard benchmarks.
We apply statistical techniques from naturallanguageprocessing to Western and Hong Kong–based English language newspaper articles that discuss the 2019–2020 Hong Kong protests of the Anti-Extradition Law Amendment...
详细信息
With the economy developing as one of the key technologies of the construction knowledge graph, entity extraction has attracted more and more researchers' attention. To understand the latest developments, this pap...
详细信息
暂无评论