Search

Serialized output training (SOT) attracts increasing attention due to its convenience and flexibility for multi-speaker automatic speech recognition (ASR). However, it is not easy to train with attention loss only. In this paper, we propose the overlapped encoding separation (EncSep) to fully utilize the benefits of the connectionist temporal classification (CTC) and attention hybrid loss. This additional separator is inserted after the encoder to extract the multi-speaker information with CTC losses. Furthermore, we propose the serialized speech information guidance SOT (GEncSep) to further utilize the separated encodings. The separated streams are concatenated to provide single-speaker information to guide attention during decoding. The experimental results on LibriMix show that the single-speaker encoding can be separated from the overlapped encoding. The CTC loss helps to improve the encoder representation under complex scenarios. GEncSep further improved performance., Comment: something wrong with input data

With the strong representational power of large language models (LLMs), generative error correction (GER) for automatic speech recognition (ASR) aims to provide semantic and phonetic refinements to address ASR errors. This work explores how LLM-based GER can enhance and expand the capabilities of Japanese language processing, presenting the first GER benchmark for Japanese ASR with 0.9-2.6k text utterances. We also introduce a new multi-pass augmented generative error correction (MPA GER) by integrating multiple system hypotheses on the input side with corrections from multiple LLMs on the output side and then merging them. To the best of our knowledge, this is the first investigation of the use of LLMs for Japanese GER, which involves second-pass language modeling on the output transcriptions generated by the ASR system (e.g., N-best hypotheses). Our experiments demonstrated performance improvement in the proposed methods of ASR quality and generalization both in SPREDS-U1-ja and CSJ data., Comment: submitted to SLT2024

Variability in medical image segmentation, arising from annotator preferences, expertise, and their choice of tools, has been well documented. While the majority of multi-annotator segmentation approaches focus on modeling annotator-specific preferences, they require annotator-segmentation correspondence. In this work, we introduce the problem of segmentation style discovery, and propose StyleSeg, a segmentation method that learns plausible, diverse, and semantically consistent segmentation styles from a corpus of image-mask pairs without any knowledge of annotator correspondence. StyleSeg consistently outperforms competing methods on four publicly available skin lesion segmentation (SLS) datasets. We also curate ISIC-MultiAnnot, the largest multi-annotator SLS dataset with annotator correspondence, and our results show a strong alignment, using our newly proposed measure AS2, between the predicted styles and annotator preferences. The code and the dataset are available at https://github.com/sfu-mial/StyleSeg., Comment: Medical Image Computing and Computer-Assisted Intervention (MICCAI) ISIC Skin Image Analysis Workshop (MICCAI ISIC) 2024; 13 pages, 2 tables, 3 figures

Recent approaches for empathetic response generation mainly focus on emotional resonance and user understanding, without considering the system's personality. Consistent personality is evident in real human expression and is important for creating trustworthy systems. To address this problem, we propose StyEmp, which aims to stylize the empathetic response generation with a consistent personality. Specifically, it incorporates a multi-grained prefix mechanism designed to capture the intricate relationship between a system's personality and its empathetic expressions. Furthermore, we introduce a personality reinforcement module that leverages contrastive learning to calibrate the generation model, ensuring that responses are both empathetic and reflective of a distinct personality. Automatic and human evaluations on the EMPATHETICDIALOGUES benchmark show that StyEmp outperforms competitive baselines in terms of both empathy and personality expressions., Comment: Accepted by the 25th Meeting of the Special Interest Group on Discourse and Dialogue (SIGDIAL 2024)

We propose a novel change-point detection method based on online Dynamic Mode Decomposition with control (ODMDwC). Leveraging ODMDwC's ability to find and track linear approximation of a non-linear system while incorporating control effects, the proposed method dynamically adapts to its changing behavior due to aging and seasonality. This approach enables the detection of changes in spatial, temporal, and spectral patterns, providing a robust solution that preserves correspondence between the score and the extent of change in the system dynamics. We formulate a truncated version of ODMDwC and utilize higher-order time-delay embeddings to mitigate noise and extract broad-band features. Our method addresses the challenges faced in industrial settings where safety-critical systems generate non-uniform data streams while requiring timely and accurate change-point detection to protect profit and life. Our results demonstrate that this method yields intuitive and improved detection results compared to the Singular-Value-Decomposition-based method. We validate our approach using synthetic and real-world data, showing its competitiveness to other approaches on complex systems' benchmark datasets. Provided guidelines for hyperparameters selection enhance our method's practical applicability., Comment: preprint under review in Elsevier

To achieve quantum computing using semiconductor spin qubits, the spin qubits must be precisely controlled. However, unexpected noise limits this precision and prevents the implementation of error correction codes. Specifically, frequency shifts have been found to suppress one-qubit gate fidelity. Although the exact source of the frequency shifts remains unclear, several experiments have indicated a relationship between the frequency shift and heating from the control signal. Based on these results, we modeled the heating effect using the low-temperature dependence of specific heat from the Debye model to simulate qubit control with frequency shifts. This model quantitatively reproduces some of the experimental results and makes it possible to simulate qubit gates with temperature-dependent frequency shifts. The simulated control fidelity of the proposed model was consistent with the experimental results, demonstrating that a gate fidelity above $99.9$% can be achieved in a hot environment. Additionally, our comparison of the two temperature-dependent frequency shift models reveals the condition for a microscopic frequency shift model in which the variance of the frequency shift is limited to a specific range. Furthermore, we investigated gate fidelity using the DRAG method and demonstrated the utility of temperature-dependent simulations., Comment: 10 pages, 9 figures

This paper introduces LLM-jp, a cross-organizational project for the research and development of Japanese large language models (LLMs). LLM-jp aims to develop open-source and strong Japanese LLMs, and as of this writing, more than 1,500 participants from academia and industry are working together for this purpose. This paper presents the background of the establishment of LLM-jp, summaries of its activities, and technical reports on the LLMs developed by LLM-jp. For the latest activities, visit https://llm-jp.nii.ac.jp/en/.

An open challenge in reinforcement learning (RL) is the effective deployment of a trained policy to new or slightly different situations as well as semantically-similar environments. We introduce Symmetry-Invariant Transformer (SiT), a scalable vision transformer (ViT) that leverages both local and global data patterns in a self-supervised manner to improve generalisation. Central to our approach is Graph Symmetric Attention, which refines the traditional self-attention mechanism to preserve graph symmetries, resulting in invariant and equivariant latent representations. We showcase SiT's superior generalization over ViTs on MiniGrid and Procgen RL benchmarks, and its sample efficiency on Atari 100k and CIFAR10., Comment: 9 main pages, accepted to ICML2024

Dynamical systems are often time-varying, whose modeling requires a function that evolves with respect to time. Recent studies such as the neural ordinary differential equation proposed a time-dependent neural network, which provides a neural network varying with respect to time. However, we claim that the architectural choice to build a time-dependent neural network significantly affects its time-awareness but still lacks sufficient validation in its current states. In this study, we conduct an in-depth analysis of the architecture of modern time-dependent neural networks. Here, we report a vulnerability of vanishing timestep embedding, which disables the time-awareness of a time-dependent neural network. Furthermore, we find that this vulnerability can also be observed in diffusion models because they employ a similar architecture that incorporates timestep embedding to discriminate between different timesteps during a diffusion process. Our analysis provides a detailed description of this phenomenon as well as several solutions to address the root cause. Through experiments on neural ordinary differential equations and diffusion models, we observed that ensuring alive time-awareness via proposed solutions boosted their performance, which implies that their current implementations lack sufficient time-dependency., Comment: 14 pages, 7 figures

Non-autoregressive (NAR) language models are known for their low latency in neural machine translation (NMT). However, a performance gap exists between NAR and autoregressive models due to the large decoding space and difficulty in capturing dependency between target words accurately. Compounding this, preparing appropriate training data for NAR models is a non-trivial task, often exacerbating exposure bias. To address these challenges, we apply reinforcement learning (RL) to Levenshtein Transformer, a representative edit-based NAR model, demonstrating that RL with self-generated data can enhance the performance of edit-based NAR models. We explore two RL approaches: stepwise reward maximization and episodic reward maximization. We discuss the respective pros and cons of these two approaches and empirically verify them. Moreover, we experimentally investigate the impact of temperature setting on performance, confirming the importance of proper temperature setting for NAR models' training., Comment: NAACL SRW 2024

We generalized a voice morphing algorithm capable of handling temporally variable, multiple-attributes, and multiple instances. The generalized morphing provides a new strategy for investigating speech diversity. However, excessive complexity and the difficulty of preparation have prevented researchers and students from enjoying its benefits. To address this issue, we introduced a set of interactive tools to make preparation and tests less cumbersome. These tools are integrated into our previously reported interactive tools as extensions. The introduction of the extended tools in lessons in graduate education was successful. Finally, we outline further extensions to explore excessively complex morphing parameter settings., Comment: 5 pages, 7 figures, submitted to Acoustical Science and Technology of Acoustical Society of Japan

This paper presents the first results of the study of high-energy electron and muon neutrino charged-current interactions in the FASER$\nu$ emulsion/tungsten detector of the FASER experiment at the LHC. A subset of the FASER$\nu$ volume, which corresponds to a target mass of 128.6~kg, was exposed to neutrinos from the LHC $pp$ collisions with a centre-of-mass energy of 13.6~TeV and an integrated luminosity of 9.5 fb$^{-1}$. Applying stringent selections requiring electrons with reconstructed energy above 200~GeV, four electron neutrino interaction candidate events are observed with an expected background of $0.025^{+0.015}_{-0.010}$, leading to a statistical significance of 5.2$\sigma$. This is the first direct observation of electron neutrino interactions at a particle collider. Eight muon neutrino interaction candidate events are also detected, with an expected background of $0.22^{+0.09}_{-0.07}$, leading to a statistical significance of 5.7$\sigma$. The signal events include neutrinos with energies in the TeV range, the highest-energy electron and muon neutrinos ever detected from an artificial source. The energy-independent part of the interaction cross section per nucleon is measured over an energy range of 560--1740 GeV (520--1760 GeV) for $\nu_e$ ($\nu_{\mu}$) to be $(1.2_{-0.7}^{+0.8}) \times 10^{-38}~\mathrm{cm}^{2}\,\mathrm{GeV}^{-1}$ ($(0.5\pm0.2) \times 10^{-38}~\mathrm{cm}^{2}\,\mathrm{GeV}^{-1}$), consistent with Standard Model predictions. These are the first measurements of neutrino interaction cross sections in those energy ranges.

This paper investigates the application of voice activity projection (VAP), a predictive turn-taking model for spoken dialogue, on multilingual data, encompassing English, Mandarin, and Japanese. The VAP model continuously predicts the upcoming voice activities of participants in dyadic dialogue, leveraging a cross-attention Transformer to capture the dynamic interplay between participants. The results show that a monolingual VAP model trained on one language does not make good predictions when applied to other languages. However, a multilingual model, trained on all three languages, demonstrates predictive performance on par with monolingual models across all languages. Further analyses show that the multilingual model has learned to discern the language of the input signal. We also analyze the sensitivity to pitch, a prosodic cue that is thought to be important for turn-taking. Finally, we compare two different audio encoders, contrastive predictive coding (CPC) pre-trained on English, with a recent model based on multilingual wav2vec 2.0 (MMS)., Comment: This paper has been accepted for presentation at The 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024) and represents the author's version of the work

This paper presents a novel approach for estimating the Koopman operator defined on a reproducing kernel Hilbert space (RKHS) and its spectra. We propose an estimation method, what we call Jet Dynamic Mode Decomposition (JetDMD), leveraging the intrinsic structure of RKHS and the geometric notion known as jets to enhance the estimation of the Koopman operator. This method refines the traditional Extended Dynamic Mode Decomposition (EDMD) in accuracy, especially in the numerical estimation of eigenvalues. This paper proves JetDMD's superiority through explicit error bounds and convergence rate for special positive definite kernels, offering a solid theoretical foundation for its performance. We also delve into the spectral analysis of the Koopman operator, proposing the notion of extended Koopman operator within a framework of rigged Hilbert space. This notion leads to a deeper understanding of estimated Koopman eigenfunctions and capturing them outside the original function space. Through the theory of rigged Hilbert space, our study provides a principled methodology to analyze the estimated spectrum and eigenfunctions of Koopman operators, and enables eigendecomposition within a rigged RKHS. We also propose a new effective method for reconstructing the dynamical system from temporally-sampled trajectory data of the dynamical system with solid theoretical guarantee. We conduct several numerical simulations using the van der Pol oscillator, the Duffing oscillator, the H\'enon map, and the Lorenz attractor, and illustrate the performance of JetDMD with clear numerical computations of eigenvalues and accurate predictions of the dynamical systems., Comment: We correct several typos. We have released the code for the numerical simulation at https://github.com/1sa014kawa/JetDMD

Adapting an automatic speech recognition (ASR) system to unseen noise environments is crucial. Integrating adapters into neural networks has emerged as a potent technique for transfer learning. This study thoroughly investigates adapter-based ASR adaptation in noisy environments. We conducted experiments using the CHiME--4 dataset. The results show that inserting the adapter in the shallow layer yields superior effectiveness, and there is no significant difference between adapting solely within the shallow layer and adapting across all layers. The simulated data helps the system to improve its performance under real noise conditions. Nonetheless, when the amount of data is the same, the real data is more effective than the simulated data. Multi-condition training is still useful for adapter training. Furthermore, integrating adapters into speech enhancement-based ASR systems yields substantial improvements.

We investigate the impact of politeness levels in prompts on the performance of large language models (LLMs). Polite language in human communications often garners more compliance and effectiveness, while rudeness can cause aversion, impacting response quality. We consider that LLMs mirror human communication traits, suggesting they align with human cultural norms. We assess the impact of politeness in prompts on LLMs across English, Chinese, and Japanese tasks. We observed that impolite prompts often result in poor performance, but overly polite language does not guarantee better outcomes. The best politeness level is different according to the language. This phenomenon suggests that LLMs not only reflect human behavior but are also influenced by language, particularly in different cultural contexts. Our findings highlight the need to factor in politeness for cross-cultural natural language processing and LLM usage.

The handling of communication breakdowns and loss of engagement is an important aspect of spoken dialogue systems, particularly for chatting systems such as attentive listening, where the user is mostly speaking. We presume that a human is best equipped to handle this task and rescue the flow of conversation. To this end, we propose a semi-autonomous system, where a remote operator can take control of an autonomous attentive listening system in real-time. In order to make human intervention easy and consistent, we introduce automatic detection of low interest and engagement to provide explicit takeover prompts to the remote operator. We implement this semi-autonomous system which detects takeover points for the operator and compare it to fully tele-operated and fully autonomous attentive listening systems. We find that the semi-autonomous system is generally perceived more positively than the autonomous system. The results suggest that identifying points of conversation when the user starts to lose interest may help us improve a fully autonomous dialogue system.

In the realm of human-AI dialogue, the facilitation of empathetic responses is important. Validation is one of the key communication techniques in psychology, which entails recognizing, understanding, and acknowledging others' emotional states, thoughts, and actions. This study introduces the first framework designed to engender empathetic dialogue with validating responses. Our approach incorporates a tripartite module system: 1) validation timing detection, 2) users' emotional state identification, and 3) validating response generation. Utilizing Japanese EmpatheticDialogues dataset - a textual-based dialogue dataset consisting of 8 emotional categories from Plutchik's wheel of emotions - the Task Adaptive Pre-Training (TAPT) BERT-based model outperforms both random baseline and the ChatGPT performance, in term of F1-score, in all modules. Further validation of our model's efficacy is confirmed in its application to the TUT Emotional Storytelling Corpus (TESC), a speech-based dialogue dataset, by surpassing both random baseline and the ChatGPT. This consistent performance across both textual and speech-based dialogues underscores the effectiveness of our framework in fostering empathetic human-AI communication., Comment: This paper has been accepted for presentation at International Workshop on Spoken Dialogue Systems Technology 2024 (IWSDS 2024)

The Forward Search Experiment (FASER) at CERN's Large Hadron Collider (LHC) has recently directly detected the first collider neutrinos. Neutrinos play an important role in all FASER analyses, either as signal or background, and it is therefore essential to understand the neutrino event rates. In this study, we update previous simulations and present prescriptions for theoretical predictions of neutrino fluxes and cross sections, together with their associated uncertainties. With these results, we discuss the potential for possible measurements that could be carried out in the coming years with the FASER neutrino data to be collected in LHC Run 3 and Run 4., Comment: 19 pages, 10 figures

Gradient-based hyperparameter optimization methods update hyperparameters using hypergradients, gradients of a meta criterion with respect to hyperparameters. Previous research used two distinct update strategies: optimizing hyperparameters using global hypergradients obtained after completing model training or local hypergradients derived after every few model updates. While global hypergradients offer reliability, their computational cost is significant; conversely, local hypergradients provide speed but are often suboptimal. In this paper, we propose *glocal* hypergradient estimation, blending "global" quality with "local" efficiency. To this end, we use the Koopman operator theory to linearize the dynamics of hypergradients so that the global hypergradients can be efficiently approximated only by using a trajectory of local hypergradients. Consequently, we can optimize hyperparameters greedily using estimated global hypergradients, achieving both reliability and efficiency simultaneously. Through numerical experiments of hyperparameter optimization, including optimization of optimizers, we demonstrate the effectiveness of the glocal hypergradient estimation.

Automatic Mean Opinion Score (MOS) prediction is employed to evaluate the quality of synthetic speech. This study extends the application of predicted MOS to the task of Fake Audio Detection (FAD), as we expect that MOS can be used to assess how close synthesized speech is to the natural human voice. We propose MOS-FAD, where MOS can be leveraged at two key points in FAD: training data selection and model fusion. In training data selection, we demonstrate that MOS enables effective filtering of samples from unbalanced datasets. In the model fusion, our results demonstrate that incorporating MOS as a gating mechanism in FAD model fusion enhances overall performance., Comment: Accepted in ICASSP2024

Audio-visual speech recognition (AVSR) is a multimodal extension of automatic speech recognition (ASR), using video as a complement to audio. In AVSR, considerable efforts have been directed at datasets for facial features such as lip-readings, while they often fall short in evaluating the image comprehension capabilities in broader contexts. In this paper, we construct SlideAVSR, an AVSR dataset using scientific paper explanation videos. SlideAVSR provides a new benchmark where models transcribe speech utterances with texts on the slides on the presentation recordings. As technical terminologies that are frequent in paper explanations are notoriously challenging to transcribe without reference texts, our SlideAVSR dataset spotlights a new aspect of AVSR problems. As a simple yet effective baseline, we propose DocWhisper, an AVSR model that can refer to textual information from slides, and confirm its effectiveness on SlideAVSR., Comment: 3rd Workshop on Advances in Language and Vision Research (ALVR 2024)

Personality recognition is useful for enhancing robots' ability to tailor user-adaptive responses, thus fostering rich human-robot interactions. One of the challenges in this task is a limited number of speakers in existing dialogue corpora, which hampers the development of robust, speaker-independent personality recognition models. Additionally, accurately modeling both the interdependencies among interlocutors and the intra-dependencies within the speaker in dialogues remains a significant issue. To address the first challenge, we introduce personality trait interpolation for speaker data augmentation. For the second, we propose heterogeneous conversational graph networks to independently capture both contextual influences and inherent personality traits. Evaluations on the RealPersonaChat corpus demonstrate our method's significant improvements over existing baselines., Comment: This paper has been accepted for presentation at International Workshop on Spoken Dialogue Systems Technology 2024 (IWSDS 2024)

A demonstration of a real-time and continuous turn-taking prediction system is presented. The system is based on a voice activity projection (VAP) model, which directly maps dialogue stereo audio to future voice activities. The VAP model includes contrastive predictive coding (CPC) and self-attention transformers, followed by a cross-attention transformer. We examine the effect of the input context audio length and demonstrate that the proposed system can operate in real-time with CPU settings, with minimal performance degradation., Comment: This paper has been accepted for presentation at International Workshop on Spoken Dialogue Systems Technology 2024 (IWSDS 2024) and represents the author's version of the work

Establishing evaluation schemes for spoken dialogue systems is important, but it can also be challenging. While subjective evaluations are commonly used in user experiments, objective evaluations are necessary for research comparison and reproducibility. To address this issue, we propose a framework for indirectly but objectively evaluating systems based on users' behaviors. In this paper, to this end, we investigate the relationship between user behaviors and subjective evaluation scores in social dialogue tasks: attentive listening, job interview, and first-meeting conversation. The results reveal that in dialogue tasks where user utterances are primary, such as attentive listening and job interview, indicators like the number of utterances and words play a significant role in evaluation. Observing disfluency also can indicate the effectiveness of formal tasks, such as job interview. On the other hand, in dialogue tasks with high interactivity, such as first-meeting conversation, behaviors related to turn-taking, like average switch pause length, become more important. These findings suggest that selecting appropriate user behaviors can provide valuable insights for objective evaluation in each social dialogue task., Comment: This paper has been accepted for presentation at International Workshop on Spoken Dialogue Systems Technology 2024 (IWSDS 2024) and represents the author's version of the work

We apply Dynamic Mode Decomposition (DMD) to Elementary Cellular Automata (ECA). Three types of DMD methods are considered and the reproducibility of the system dynamics and Koopman eigenvalues from observed time series are investigated. While standard DMD fails to reproduce the system dynamics and Koopman eigenvalues associated with a given periodic orbit in some cases, Hankel DMD with delay-embedded time series improves reproducibility. However, Hankel DMD can still fail to reproduce all the Koopman eigenvalues in specific cases. We propose an Extended DMD method for ECA that uses nonlinearly transformed time series with discretized Walsh functions and show that it can completely reproduce the dynamics and Koopman eigenvalues. Linear-algebraic backgrounds for the reproducibility of the system dynamics and Koopman eigenvalues are also discussed., Comment: 20 pages, 10 figures

Dynamic mode (DM) decomposition decomposes spatiotemporal signals into basic oscillatory components (DMs). DMs can improve the accuracy of neural decoding when used with the nonlinear Grassmann kernel, compared to conventional power features. However, such kernel-based machine learning algorithms have three limitations: large computational time preventing real-time application, incompatibility with non-kernel algorithms, and low interpretability. Here, we propose a mapping function corresponding to the Grassmann kernel that explicitly transforms DMs into spatial DM (sDM) features, which can be used in any machine learning algorithm. Using electrocorticographic signals recorded during various movement and visual perception tasks, the sDM features were shown to improve the decoding accuracy and computational time compared to conventional methods. Furthermore, the components of the sDM features informative for decoding showed similar characteristics to the high-$\gamma$ power of the signals, but with higher trial-to-trial reproducibility. The proposed sDM features enable fast, accurate, and interpretable neural decoding.

The effective detection of evidence of financial anomalies requires collaboration among multiple entities who own a diverse set of data, such as a payment network system (PNS) and its partner banks. Trust among these financial institutions is limited by regulation and competition. Federated learning (FL) enables entities to collaboratively train a model when data is either vertically or horizontally partitioned across the entities. However, in real-world financial anomaly detection scenarios, the data is partitioned both vertically and horizontally and hence it is not possible to use existing FL approaches in a plug-and-play manner. Our novel solution, PV4FAD, combines fully homomorphic encryption (HE), secure multi-party computation (SMPC), differential privacy (DP), and randomization techniques to balance privacy and accuracy during training and to prevent inference threats at model deployment time. Our solution provides input privacy through HE and SMPC, and output privacy against inference time attacks through DP. Specifically, we show that, in the honest-but-curious threat model, banks do not learn any sensitive features about PNS transactions, and the PNS does not learn any information about the banks' dataset but only learns prediction labels. We also develop and analyze a DP mechanism to protect output privacy during inference. Our solution generates high-utility models by significantly reducing the per-bank noise level while satisfying distributed DP. To ensure high accuracy, our approach produces an ensemble model, in particular, a random forest. This enables us to take advantage of the well-known properties of ensembles to reduce variance and increase accuracy. Our solution won second prize in the first phase of the U.S. Privacy Enhancing Technologies (PETs) Prize Challenge., Comment: Prize Winner in the U.S. Privacy Enhancing Technologies (PETs) Prize Challenge

Automatic evaluation of text generation is essential for improving the accuracy of generation tasks. In light of the current trend towards increasingly larger decoder-based language models, we investigate automatic evaluation methods based on such models for text generation. This paper compares various methods, including tuning with encoder-based models and large language models under equal conditions, on two different tasks, machine translation evaluation and semantic textual similarity, in two languages, Japanese and English. Experimental results show that compared to the tuned encoder-based models, the tuned decoder-based models perform poorly. The analysis of the causes for this suggests that the decoder-based models focus on surface word sequences and do not capture meaning. It is also revealed that in-context learning of very large decoder-based models such as ChatGPT makes it difficult to identify fine-grained semantic differences., Comment: Accepted to IJCNLP-AACL 2023 SRW

Despite the remarkable progress in natural language understanding with pretrained Transformers, neural language models often do not handle commonsense knowledge well. Toward commonsense-aware models, there have been attempts to obtain knowledge, ranging from automatic acquisition to crowdsourcing. However, it is difficult to obtain a high-quality knowledge base at a low cost, especially from scratch. In this paper, we propose PHALM, a method of building a knowledge graph from scratch, by prompting both crowdworkers and a large language model (LLM). We used this method to build a Japanese event knowledge graph and trained Japanese commonsense generation models. Experimental results revealed the acceptability of the built graph and inferences generated by the trained models. We also report the difference in prompting humans and an LLM. Our code, data, and models are available at github.com/nlp-waseda/comet-atomic-ja.

Sound event localization and detection (SELD) systems estimate direction-of-arrival (DOA) and temporal activation for sets of target classes. Neural network (NN)-based SELD systems have performed well in various sets of target classes, but they only output the DOA and temporal activation of preset classes trained before inference. To customize target classes after training, we tackle zero- and few-shot SELD tasks, in which we set new classes with a text sample or a few audio samples. While zero-shot sound classification tasks are achievable by embedding from contrastive language-audio pretraining (CLAP), zero-shot SELD tasks require assigning an activity and a DOA to each embedding, especially in overlapping cases. To tackle the assignment problem in overlapping cases, we propose an embed-ACCDOA model, which is trained to output track-wise CLAP embedding and corresponding activity-coupled Cartesian direction-of-arrival (ACCDOA). In our experimental evaluations on zero- and few-shot SELD tasks, the embed-ACCDOA model showed better location-dependent scores than a straightforward combination of the CLAP audio encoder and a DOA estimation model. Moreover, the proposed combination of the embed-ACCDOA model and CLAP audio encoder with zero- or few-shot samples performed comparably to an official baseline system trained with complete train data in an evaluation dataset., Comment: 5 pages, 4 figures, accepted for publication in IEEE ICASSP 2024