Fine-tuning-based unlearning methods prevail for erasing targeted harmful, sensitive, or copyrighted information within large language models while preserving overall capabilities. However, the true effectiveness of the methods is unclear. In this paper, we delve into the limitations of fine-tuning-based unlearning through activation patching and parameter restoration experiments. Our findings reveal that these methods alter the model’s knowledge retrieval process, rather than genuinely erasing the problematic knowledge embedded in the model parameters. Furthermore, behavioral tests demonstrate that the unlearning mechanisms inevitably impact the global behavior of the models, affecting unrelated knowledge or capabilities. Our work advocates the development of more resilient unlearning techniques for truly erasing knowledge.

Recently it has been shown that deep learning models for NLP tasks are prone to attacks that can even reconstruct the verbatim training texts. To prevent privacy leakage, researchers have investigated word-level perturbations, relying on the formal guarantees of differential privacy (DP) in the embedding space. However, many existing approaches either achieve unsatisfactory performance in the high privacy regime when using the Laplacian or Gaussian mechanism, or resort to weaker relaxations of DP that are inferior to the canonical DP in terms of privacy strength. This raises the question of whether a new method for private word embedding can be designed to overcome these limitations. In this paper, we propose a novel private embedding method called the high dimensional truncated Laplacian mechanism. Specifically, we introduce a non-trivial extension of the truncated Laplacian mechanism, which was previously only investigated in one-dimensional space cases. Theoretically, we show that our method has a lower variance compared to the previous private word embedding methods. To further validate its effectiveness, we conduct comprehensive experiments on private embedding and downstream tasks using three datasets. Remarkably, even in the high privacy regime, our approach only incurs a slight decrease in utility compared to the non-private scenario.

Recent developments in deep learning have led to great success in various natural language processing (NLP) tasks. However, these applications may involve data that contain sensitive information. Therefore, how to achieve good performance while also protecting the privacy of sensitive data is a crucial challenge in NLP. To preserve privacy, Differential Privacy (DP), which can prevent reconstruction attacks and protect against potential side knowledge, is becoming a de facto technique for private data analysis. In recent years, NLP in DP models (DP-NLP) has been studied from different perspectives, which deserves a comprehensive review. In this paper, we provide the first systematic review of recent advances in DP deep learning models in NLP. In particular, we first discuss some differences and additional challenges of DP-NLP compared with the standard DP deep learning. Then, we investigate some existing work on DP-NLP andpresent its recent developments from three aspects: gradient perturbation based methods, embedding vector perturbation based methods, and ensemble model based methods. We also discuss some challenges and future directions.

Antonyms vs synonyms distinction is a core challenge in lexico-semantic analysis and automated lexical resource construction. These pairs share a similar distributional context which makes it harder to distinguish them. Leading research in this regard attempts to capture the properties of the relation pairs, i.e., symmetry, transitivity, and trans-transitivity. However, the inability of existing research to appropriately model the relation-specific properties limits their end performance. In this paper, we propose InterlaCed Encoder NETworks (i.e., ICE-NET) for antonym vs synonym distinction, that aim to capture and model the relation-specific properties of the antonyms and synonyms pairs in order to perform the classification task in a performance-enhanced manner. Experimental evaluation using the benchmark datasets shows that ICE-NET outperforms the existing research by a relative score of upto 1.8% in F1-measure.

Autonomous artificial intelligence (AI) agents have emerged as promising protocols for automatically understanding the language-based environment, particularly with the exponential development of large language models (LLMs). However, a fine-grained, comprehensive understanding of multimodal environments remains under-explored. This work designs an autonomous workflow tailored for integrating AI agents seamlessly into extended reality (XR) applications for fine-grained training. We present a demonstration of a multimodal fine-grained training assistant for LEGO brick assembly in a pilot XR environment. Specifically, we design a cerebral language agent that integrates LLM with memory, planning, and interaction with XR tools and a vision-language agent, enabling agents to decide their actions based on past experiences. Furthermore, we introduce LEGO-MRTA, a multimodal fine-grained assembly dialogue dataset synthesized automatically in the workflow served by a commercial LLM. This dataset comprises multimodal instruction manuals, conversations, XR responses, and vision question answering. Last, we present several prevailing open-resource LLMs as benchmarks, assessing their performance with and without fine-tuning on the proposed dataset. We anticipate that the broader impact of this workflow will advance the development of smarter assistants for seamless user interaction in XR environments, fostering research in both AI and HCI communities.

During crisis situations, observations of other people’s behaviors often play an essential role in a person’s decision-making. For example, a person might evacuate before a hurricane only if everyone else in the neighborhood does so. Conversely, a person might stay if no one else is leaving. Such observations are called social cues. Social cues are important for understanding people’s response to crises, so recognizing them can help inform the decisions of government officials and emergency responders. In this paper, we propose the first NLP task to categorize social cues in social media posts during crisis situations. We introduce a manually annotated dataset of 6,000 tweets, labeled with respect to eight social cue categories. We also present experimental results of several classification models, which show that some types of social cues can be recognized reasonably well, but overall this task is challenging for NLP systems. We further present error analyses to identify specific types of mistakes and promising directions for future research on this task.

End-to-end multimodal aspect-based sentiment analysis (MABSA) combines multimodal aspect terms extraction (MATE) with multimodal aspect sentiment classification (MASC), aiming to simultaneously extract aspect words and classify the sentiment polarity of each aspect. However, existing MABSA methods have overlooked two issues: (i) They only focus on fusing image regional information and textual words for two subtasks of MABSA. Whereas, MATE subtask relies more on global image information to assist in obtaining the quantity and attributes of aspects. Ignoring the integration with global information may affect the performance of MABSA methods. (ii) They fail to take advantage of target information. Nevertheless, the fine-grained details of targets are important for classifying sentiments of aspects. To solve these problems, we propose a Target-oriented Multi-grained Fusion Network(TMFN). It fuses text information with global coarse-grained image information for MATE subtask and with fine-grained image information for MASC subtask. In addition, a target-oriented feature alignment (TOFA) module is designed to enhance target-related information in image features with target details. In such a way, image features will contain more target emotional-related information which is beneficial to sentiment classification. Extensive experiments show that our method outperforms state-of-the-art methods on two benchmark datasets.

Automated construction of bi-lingual dictionaries using monolingual embedding spaces is a core challenge in machine translation. The end performance of these dictionaries relies on the geometric similarity of individual spaces, i.e., their degree of isomorphism. Existing attempts aimed at controlling the relative isomorphism of different spaces fail to incorporate the impact of lexically different but semantically related words in the training objective. To address this, we propose GRI that combines the distributional training objectives with attentive graph convolutions to unanimously consider the impact of lexical variations of semantically similar words required to define/compute the relative isomorphism of multiple spaces. Exper imental evaluation shows that GRI outperforms the existing research by improving the average P@1 by a relative score of upto 63.6%.

With the rapid progress of Large language models (LLMs) and the huge amount of text they generate, it becomes impractical to manually distinguish whether a text is machine-generated. The growing use of LLMs in social media and education, prompts us to develop methods to detect machine-generated text, preventing malicious use such as plagiarism, misinformation, and propaganda. In this paper, we introduce two novel zero-shot methods for detecting machine-generated text by leveraging the Log-Rank information. One is called DetectLLM-LRR, which is fast and efficient, and the other is called DetectLLM-NPR, which is more accurate, but slower due to the need for perturbations. Our experiments on three datasets and seven language models show that our proposed methods improve over the state of the art by 3.9 and 1.75 AUROC points absolute. Moreover, DetectLLM-NPR needs fewer perturbations than previous work to achieve the same level of performance, which makes it more practical for real-world use. We also investigate the efficiency-performance trade-off based on users’ preference for these two measures and provide intuition for using them in practice effectively. We release the data and the code of both methods in https://github.com/mbzuai-nlp/DetectLLM.

Bilingual Lexical Induction (BLI) is a core challenge in NLP, it relies on the relative isomorphism of individual embedding spaces. Existing attempts aimed at controlling the relative isomorphism of different embedding spaces fail to incorporate the impact of semantically related words in the model training objective. To address this, we propose GARI that combines the distributional training objectives with multiple isomorphism losses guided by the graph attention network. GARI considers the impact of semantical variations of words in order to define the relative isomorphism of the embedding spaces. Experimental evaluation using the Arabic language data set shows that GARI outperforms the existing research by improving the average P@1 by a relative score of up to 40.95% and 76.80% for in-domain and domain mismatch settings respectively.

Evaluation metrics shine the light on the best models and thus strongly influence the research directions, such as the recently developed dialogue metrics USR, FED, and GRADE. However, most current metrics evaluate the dialogue data as isolated and static because they only focus on a single quality or several qualities. To mitigate the problem, this paper proposes an interpretable, multi-faceted, and controllable framework IM^2 (Interpretable and Multi-category Integrated Metric) to combine a large number of metrics which are good at measuring different qualities. The IM^2 framework first divides current popular dialogue qualities into different categories and then applies or proposes dialogue metrics to measure the qualities within each category and finally generates an overall IM^2 score. An initial version of IM^2 was submitted to the AAAI 2022 Track5.1@DSTC10 challenge and took the 2^nd place on both of the development and test leaderboard. After the competition, we develop more metrics and improve the performance of our model. We compare IM^2 with other 13 current dialogue metrics and experimental results show that IM^2 correlates more strongly with human judgments than any of them on each evaluated dataset.

Chinese spelling correction (CSC) is a task to detect and correct spelling errors in texts. CSC is essentially a linguistic problem, thus the ability of language understanding is crucial to this task. In this paper, we propose a Pre-trained masked Language model with Misspelled knowledgE (PLOME) for CSC, which jointly learns how to understand language and correct spelling errors. To this end, PLOME masks the chosen tokens with similar characters according to a confusion set rather than the fixed token “[MASK]” as in BERT. Besides character prediction, PLOME also introduces pronunciation prediction to learn the misspelled knowledge on phonic level. Moreover, phonological and visual similarity knowledge is important to this task. PLOME utilizes GRU networks to model such knowledge based on characters’ phonics and strokes. Experiments are conducted on widely used benchmarks. Our method achieves superior performance against state-of-the-art approaches by a remarkable margin. We release the source code and pre-trained model for further use by the community (https://github.com/liushulinle/PLOME).

Hierarchical Text Classification (HTC) is a challenging task that categorizes a textual description within a taxonomic hierarchy. Most of the existing methods focus on modeling the text. Recently, researchers attempt to model the class representations with some resources (e.g., external dictionaries). However, the concept shared among classes which is a kind of domain-specific and fine-grained information has been ignored in previous work. In this paper, we propose a novel concept-based label embedding method that can explicitly represent the concept and model the sharing mechanism among classes for the hierarchical text classification. Experimental results on two widely used datasets prove that the proposed model outperforms several state-of-the-art methods. We release our complementary resources (concepts and definitions of classes) for these two datasets to benefit the research on HTC.

We introduce Texar, an open-source toolkit aiming to support the broad set of text generation tasks that transform any inputs into natural language, such as machine translation, summarization, dialog, content manipulation, and so forth. With the design goals of modularity, versatility, and extensibility in mind, Texar extracts common patterns underlying the diverse tasks and methodologies, creates a library of highly reusable modules and functionalities, and allows arbitrary model architectures and algorithmic paradigms. In Texar, model architecture, inference, and learning processes are properly decomposed. Modules at a high concept level can be freely assembled or plugged in/swapped out. Texar is thus particularly suitable for researchers and practitioners to do fast prototyping and experimentation. The versatile toolkit also fosters technique sharing across different text generation tasks. Texar supports both TensorFlow and PyTorch, and is released under Apache License 2.0 at https://www.texar.io.

Recent approaches to cross-lingual word embedding have generally been based on linear transformations between the sets of embedding vectors in the two languages. In this paper, we propose an approach that instead expresses the two monolingual embedding spaces as probability densities defined by a Gaussian mixture model, and matches the two densities using a method called normalizing flow. The method requires no explicit supervision, and can be learned with only a seed dictionary of words that have identical strings. We argue that this formulation has several intuitively attractive properties, particularly with the respect to improving robustness and generalization to mappings between difficult language pairs or word pairs. On a benchmark data set of bilingual lexicon induction and cross-lingual word similarity, our approach can achieve competitive or superior performance compared to state-of-the-art published results, with particularly strong results being found on etymologically distant and/or morphologically rich languages.

Messages in human conversations inherently convey emotions. The task of detecting emotions in textual conversations leads to a wide range of applications such as opinion mining in social networks. However, enabling machines to analyze emotions in conversations is challenging, partly because humans often rely on the context and commonsense knowledge to express emotions. In this paper, we address these challenges by proposing a Knowledge-Enriched Transformer (KET), where contextual utterances are interpreted using hierarchical self-attention and external commonsense knowledge is dynamically leveraged using a context-aware affective graph attention mechanism. Experiments on multiple textual conversation datasets demonstrate that both context and commonsense knowledge are consistently beneficial to the emotion detection performance. In addition, the experimental results show that our KET model outperforms the state-of-the-art models on most of the tested datasets in F1 score.

We introduce Texar, an open-source toolkit aiming to support the broad set of text generation tasks. Different from many existing toolkits that are specialized for specific applications (e.g., neural machine translation), Texar is designed to be highly flexible and versatile. This is achieved by abstracting the common patterns underlying the diverse tasks and methodologies, creating a library of highly reusable modules and functionalities, and enabling arbitrary model architectures and various algorithmic paradigms. The features make Texar particularly suitable for technique sharing and generalization across different text generation applications. The toolkit emphasizes heavily on extensibility and modularized system design, so that components can be freely plugged in or swapped out. We conduct extensive experiments and case studies to demonstrate the use and advantage of the toolkit.

Question answering systems deteriorate dramatically in the presence of adversarial sentences in articles. According to Jia and Liang (2017), the single BiDAF system (Seo et al., 2016) only achieves an F1 score of 4.8 on the ADDANY adversarial dataset. In this paper, we present a method to tackle this problem via answer sentence selection. Given a paragraph of an article and a corresponding query, instead of directly feeding the whole paragraph to the single BiDAF system, a sentence that most likely contains the answer to the query is first selected, which is done via a deep neural network based on TreeLSTM (Tai et al., 2015). Experiments on ADDANY adversarial dataset validate the effectiveness of our method. The F1 score has been improved to 52.3.

We propose simple and flexible training and decoding methods for influencing output style and topic in neural encoder-decoder based language generation. This capability is desirable in a variety of applications, including conversational systems, where successful agents need to produce language in a specific style and generate responses steered by a human puppeteer or external knowledge. We decompose the neural generation process into empirically easier sub-problems: a faithfulness model and a decoding method based on selective-sampling. We also describe training and sampling algorithms that bias the generation process with a specific language style restriction, or a topic restriction. Human evaluation results show that our proposed methods are able to to restrict style and topic without degrading output quality in conversational tasks.

The Language Application (LAPPS) Grid project is establishing a framework that enables language service discovery, composition, and reuse and promotes sustainability, manageability, usability, and interoperability of natural language Processing (NLP) components. It is based on the service-oriented architecture (SOA), a more recent, web-oriented version of the “pipeline” architecture that has long been used in NLP for sequencing loosely-coupled linguistic analyses. The LAPPS Grid provides access to basic NLP processing tools and resources and enables pipelining such tools to create custom NLP applications, as well as composite services such as question answering and machine translation together with language resources such as mono- and multi-lingual corpora and lexicons that support NLP. The transformative aspect of the LAPPS Grid is that it orchestrates access to and deployment of language resources and processing functions available from servers around the globe and enables users to add their own language resources, services, and even service grids to satisfy their particular needs.