Large language models (LLMs) have made significant advancements in various natural language processing tasks, including question answering (QA) tasks. While incorporating new information with the retrieval of relevant passages is a promising way to improve QA with LLMs, the existing methods often require additional fine-tuning which becomes infeasible with recent LLMs. Augmenting retrieved passages via prompting has the potential to address this limitation, but this direction has been limitedly explored. To this end, we design a simple yet effective framework to enhance open-domain QA (ODQA) with LLMs, based on the summarized retrieval (SuRe). SuRe helps LLMs predict more accurate answers for a given question, which are well-supported by the summarized retrieval that could be viewed as an explicit rationale extracted from the retrieved passages. Specifically, SuRe first constructs summaries of the retrieved passages for each of the multiple answer candidates. Then, SuRe confirms the most plausible answer from the candidate set by evaluating the validity and ranking of the generated summaries. Experimental results on diverse ODQA benchmarks demonstrate the superiority of SuRe, with improvements of up to 4.6\% in exact match (EM) and 4.0\% in F1 score over standard prompting approaches. SuRe also can be integrated with a broad range of retrieval methods and LLMs. Finally, the generated summaries from SuRe show additional advantages to measure the importance of retrieved passages and serve as more preferred rationales by models and humans.

Through in-context learning (ICL), large-scale language models are effective few-shot learners without additional model fine-tuning. However, the ICL performance does not scale well with the number of available training sample as it is limited by the inherent input length constraint of the underlying language model. Meanwhile, many studies have revealed that language models are also powerful feature extractors, allowing them to be utilized in a black-box manner and enabling the linear probing paradigm, where lightweight discriminators are trained on top of the pre-extracted input representations. This paper proposes prompt-augmented linear probing (PALP), a hybrid of linear probing and ICL, which leverages the best of both worlds. PALP inherits the scalability of linear probing and the capability of enforcing language models to derive more meaningful representations via tailoring input into a more conceivable form. Throughout in-depth investigations on various datasets, we verified that PALP significantly closes the gap between ICL in the data-hungry scenario and fine-tuning in the data-abundant scenario with little training overhead, potentially making PALP a strong alternative in a black-box scenario.

Deokjae Lee, JunYeong Lee, Jung-Woo Ha, Jin-Hwa Kim, Sang-Woo Lee, Hwaran Lee, Hyun Oh Song. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

Kyuyong Shin, Hanock Kwak, Wonjae Kim, Jisu Jeong, Seungjae Jung, Kyungmin Kim, Jung-Woo Ha, Sang-Woo Lee. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

In this paper, we present a beam tracking algorithm based on the deep Q-network (DQN) for mobile millimeter-wave (mmWave) communications. The proposed algorithm determines the receive beam angle from the received signals without knowing the channel model and dynamics. It uses the received signals of the current and previous time slots to design the state and reward of the DQN. Our goal is to maximize the signal-to-noise ratio by the actions of the designed DQN. A significant computational complexity reduction is achieved since the receiver does not need to run complicated signal processing algorithms once the DQN is properly trained. Therefore a practical implementation of mmWave beam tracking with a very large number of antennas under harsh mobile environments becomes feasible. Through the extensive simulations, we verified the performance of the proposed algorithm and demonstrated robustness to the system uncertainty and low computational complexity in comparison with particle filter and the Q-learning.

While Transformers have had significant success in paragraph generation, they treat sentences as linear sequences of tokens and often neglect their hierarchical information. Prior work has shown that decomposing the levels of granularity (e.g., word, phrase, or sentence) for input tokens has produced substantial improvements, suggesting the possibility of enhancing Transformers via more fine-grained modeling of granularity. In this work, we present continuous decomposition of granularity for neural paraphrase generation (C-DNPG): an advanced extension of multi-head self-attention with: 1) a granularity head that automatically infers the hierarchical structure of a sentence by neurally estimating the granularity level of each input token; and 2) two novel attention masks, namely, granularity resonance and granularity scope, to efficiently encode granularity into attention. Experiments on two benchmarks, including Quora question pairs and Twitter URLs have shown that C-DNPG outperforms baseline models by a significant margin. Qualitative analysis reveals that C-DNPG indeed captures fine-grained levels of granularity with effectiveness.

Metadata attributes (e.g., user and product IDs from reviews) can be incorporated as additional inputs to neural-based NLP models, by expanding the architecture of the models to improve performance. However, recent models rely on pretrained language models (PLMs), in which previously used techniques for attribute injection are either nontrivial or cost-ineffective. In this paper, we introduce a benchmark for evaluating attribute injection models, which comprises eight datasets across a diverse range of tasks and domains and six synthetically sparsified ones. We also propose a lightweight and memory-efficient method to inject attributes into PLMs. We extend adapters, i.e. tiny plug-in feed-forward modules, to include attributes both independently of or jointly with the text. We use approximation techniques to parameterize the model efficiently for domains with large attribute vocabularies, and training mechanisms to handle multi-labeled and sparse attributes. Extensive experiments and analyses show that our method outperforms previous attribute injection methods and achieves state-of-the-art performance on all datasets.

Kang Min Yoo, Junyeob Kim, Hyuhng Joon Kim, Hyunsoo Cho, Hwiyeol Jo, Sang-Woo Lee, Sang-goo Lee, Taeuk Kim. Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing. 2022.

Sanghwan Bae, Donghyun Kwak, Sungdong Kim, Donghoon Ham, Soyoung Kang, Sang-Woo Lee, Woomyoung Park. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Seongjin Shin, Sang-Woo Lee, Hwijeen Ahn, Sungdong Kim, HyoungSeok Kim, Boseop Kim, Kyunghyun Cho, Gichang Lee, Woomyoung Park, Jung-Woo Ha, Nako Sung. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Text-to-image generation and image captioning are recently emerged as a new experimental paradigm to assess machine intelligence. They predict continuous quantity accompanied by their sampling techniques in the generation, making evaluation complicated and intractable to get marginal distributions. Based on a recent trend that multimodal generative evaluations exploit a vison-and-language pre-trained model, we propose the negative Gaussian cross-mutual information using the CLIP features as a unified metric, coined by Mutual Information Divergence (MID). To validate, we extensively compare it with competing metrics using carefully-generated or human-annotated judgments in text-to-image generation and image captioning tasks. The proposed MID significantly outperforms the competitive methods by having consistency across benchmarks, sample parsimony, and robustness toward the exploited CLIP model. We look forward to seeing the underrepresented implications of the Gaussian cross-mutual information in multimodal representation learning and future works based on this novel proposition.

Sungdong Kim, Minsuk Chang, Sang-Woo Lee. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Boseop Kim, HyoungSeok Kim, Sang-Woo Lee, Gichang Lee, Donghyun Kwak, Jeon Dong Hyeon, Sunghyun Park, Sungju Kim, Seonhoon Kim, Dongpil Seo, Heungsub Lee, Minyoung Jeong, Sungjae Lee, Minsub Kim, Suk Hyun Ko, Seokhun Kim, Taeyong Park, Jinuk Kim, Soyoung Kang, Na-Hyeon Ryu, Kang Min Yoo, Minsuk Chang, Soobin Suh, Sookyo In, Jinseong Park, Kyungduk Kim, Hiun Kim, Jisu Jeong, Yong Goo Yeo, Donghoon Ham, Dongju Park, Min Young Lee, Jaewook Kang, Inho Kang, Jung-Woo Ha, Woomyoung Park, Nako Sung. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. 2021.

Language model pre-training has shown promising results in various downstream tasks. In this context, we introduce a cross-modal pre-trained language model, called Speech-Text BERT (ST-BERT), to tackle end-to-end spoken language understanding (E2E SLU) tasks. Taking phoneme posterior and subword-level text as an input, ST-BERT learns a contextualized cross-modal alignment via our two proposed pre-training tasks: Cross-modal Masked Language Modeling (CM-MLM) and Cross-modal Conditioned Language Modeling (CM-CLM). Experimental results on three benchmarks present that our approach is effective for various SLU datasets and shows a surprisingly marginal performance degradation even when 1% of the training data are available. Also, our method shows further SLU performance gain via domain-adaptive pre-training with domain-specific speech-text pair data.

Recent works in dialogue state tracking (DST) focus on an open vocabulary-based setting to resolve scalability and generalization issues of the predefined ontology-based approaches.However, they are inefficient in that they predict the dialogue state at every turn from scratch.Here, we consider dialogue state as an explicit fixed-sized memory and propose a selectively overwriting mechanism for more efficient DST.This mechanism consists of two steps: (1) predicting state operation on each of the memory slots, and (2) overwriting the memory with new values, of which only a few are generated according to the predicted state operations.Our method decomposes DST into two sub-tasks and guides the decoder to focus only on one of the tasks, thus reducing the burden of the decoder.This enhances the effectiveness of training and DST performance.Our SOM-DST (Selectively Overwriting Memory for Dialogue State Tracking) model achieves state-of-theart joint goal accuracy with 51.72% in Mul-tiWOZ 2.0 and 53.01% in MultiWOZ 2.1 in an open vocabulary-based DST setting.In addition, we analyze the accuracy gaps between the current and the ground truth-given situations and suggest that it is a promising direction to improve state operation prediction to boost the DST performance. 1

Automatic speech recognition (ASR) via call is essential for various applications, including AI for contact center (AICC) services. Despite the advancement of ASR, however, most publicly available call-based speech corpora such as Switchboard are old-fashioned. Also, most existing call corpora are in English and mainly focus on open domain dialog or general scenarios such as audiobooks. Here we introduce a new large-scale Korean call-based speech corpus under a goal-oriented dialog scenario from more than 11,000 people, i.e., ClovaCall corpus. ClovaCall includes approximately 60,000 pairs of a short sentence and its corresponding spoken utterance in a restaurant reservation domain. We validate the effectiveness of our dataset with intensive experiments using two standard ASR models. Furthermore, we release our ClovaCall dataset and baseline source codes to be available via https://github.com/ClovaAI/ClovaCall. Copyright © 2020 ISCA

Goal-oriented dialog has been given attention due to its numerous applications in artificial intelligence. Goal-oriented dialogue tasks occur when a questioner asks an action-oriented question and an answerer responds with the intent of letting the questioner know a correct action to take. To ask the adequate question, deep learning and reinforcement learning have been recently applied. However, these approaches struggle to find a competent recurrent neural questioner, owing to the complexity of learning a series of sentences. Motivated by theory of mind, we propose "Answerer in Questioner's Mind" (AQM), a novel information theoretic algorithm for goal-oriented dialog. With AQM, a questioner asks and infers based on an approximated probabilistic model of the answerer. The questioner figures out the answerer’s intention via selecting a plausible question by explicitly calculating the information gain of the candidate intentions and possible answers to each question. We test our framework on two goal-oriented visual dialog tasks: "MNIST Counting Dialog" and "GuessWhat?!". In our experiments, AQM outperforms comparative algorithms by a large margin.

Catastrophic forgetting is a problem of neural networks that loses the information of the first task after training the second task. Here, we propose a method, i.e. incremental moment matching (IMM), to resolve this problem. IMM incrementally matches the moment of the posterior distribution of the neural network which is trained on the first and the second task, respectively. To make the search space of posterior parameter smooth, the IMM procedure is complemented by various transfer learning techniques including weight transfer, L2-norm of the old and the new parameter, and a variant of dropout with the old parameter. We analyze our approach on a variety of datasets including the MNIST, CIFAR-10, Caltech-UCSD-Birds, and Lifelog datasets. The experimental results show that IMM achieves state-of-the-art performance by balancing the information between an old and a new network.

Deep neural networks continue to advance the state-of-the-art of image recognition tasks with various methods. However, applications of these methods to multimodality remain limited. We present Multimodal Residual Networks (MRN) for the multimodal residual learning of visual question-answering, which extends the idea of the deep residual learning. Unlike the deep residual learning, MRN effectively learns the joint representation from visual and language information. The main idea is to use element-wise multiplication for the joint residual mappings exploiting the residual learning of the attentional models in recent studies. Various alternative models introduced by multimodality are explored based on our study. We achieve the state-of-the-art results on the Visual QA dataset for both Open-Ended and Multiple-Choice tasks. Moreover, we introduce a novel method to visualize the attention effect of the joint representations for each learning block using back-propagation algorithm, even though the visual features are collapsed without spatial information.

This paper presents a multihop range-free localization algorithm that tolerates network anisotropy with a small number of anchors. A detoured path detection is proposed which measures the deviation in the hop count between the direct and shortest paths of a node pair. A novel distance estimation method is introduced to approximate the shortest path based on the path deviation and to estimate their Euclidean distance by taking into account the extent of the detour of the approximate shortest path. Compared to other range-free localization algorithms, the proposed algorithm requires fewer anchors while achieving higher localization accuracy in anisotropic networks. We demonstrated its superiority over existing range-free localization algorithms through extensive computer simulations.

This paper presents a nearest neighbor selection algorithm for real-time WiFi fingerprinting positioning with the assist of inertial measurement unit (IMU) measurements. The WiFi fingerprinting positioning using received signal strength (RSS) measurements suffers from the RSS variation problem. Due to this problem, reference points that are irrelevant to the user's position are selected, and the positioning accuracy decreases. To overcome the RSS variation problem, we propose an IMU-assisted nearest neighbor selection algorithm that filters out irrelevant reference points based on the position prediction with IMU measurements. The proposed algorithm was evaluated and compared with the conventional ii-nearest neighbors (KNN) selection and the IMU-based dead-reckoning positioning in a real indoor environment. The experimental results showed that the average positioning error of the proposed algorithm was 2.41 m, whereas those of the KNN-based fingerprinting algorithm and the IMU-based dead-reckoning positioning were 3.57 m and 15.27 m.

Due to the received signal strength (RSS) variation, WiFi indoor positioning techniques using RSS have difficulties to provide good location estimates. To mitigate the effect of the RSS variation, this paper presents a Kalman filter-based positioning algorithm that is combined with pedestrian dead reckoning and RSS-based fingerprinting positioning. The RSS recovery and clustering methods are also introduced to enhance the accuracy of the fingerprinting positioning. Unlike other existing algorithms, the proposed algorithm estimates biases accumulated in RSS measurements based on the recursive least square estimation and removes them from the measurements. Reference points are effectively selected with clustering using the recovered RSS measurements. Hence, a more accurate location estimate can be obtained in the existence of the RSS variation. The proposed algorithm is implemented into an Android-based smartphone for test.

This paper presents a multihop range-free localization algorithm to enhance the localization accuracy in anisotropic networks with a small number of anchors. We derive potential locations of a normal node and corresponding probabilities in terms of the average one-hop internodal distance, the average hop progress and the hop counts to anchors. A novel distance estimation based on the potential locations is proposed to tolerate network anisotropy from nonuniform node deployments, irregular regions, and irregular radio simultaneously. Compared to other range-free algorithms, the proposed algorithm requires fewer anchors while achieving higher localization accuracy. The superiority of the proposed algorithm against other algorithms is demonstrated through computer simulations.