Juntae Kim

Juntae Kim, Eunjung Cho, Dongbin Na

The recent program development industries have required problem-solving abilities for engineers, especially application developers. However, AI-based education systems to help solve computer algorithm problems have not yet attracted attention, while most big tech companies require the ability to solve algorithm problems including Google, Meta, and Amazon. The most useful guide to solving algorithm problems might be guessing the category (tag) of the facing problems. Therefore, our study addresses the task of predicting the algorithm tag as a useful tool for engineers and developers. Moreover, we also consider predicting the difficulty levels of algorithm problems, which can be used as useful guidance to calculate the required time to solve that problem. In this paper, we present a real-world algorithm problem multi-task dataset, AMT, by mainly collecting problem samples from the most famous and large competitive programming website Codeforces. To the best of our knowledge, our proposed dataset is the most large-scale dataset for predicting algorithm tags compared to previous studies. Moreover, our work is the first to address predicting the difficulty levels of algorithm problems. We present a deep learning-based novel method for simultaneously predicting algorithm tags and the difficulty levels of an algorithm problem given. All datasets and source codes are available at https://github.com/sronger/PSG_Predicting_Algorithm_Tags_and_Difficulty.

Youngin Cho, Junsoo Lee, Soyoung Yang et al.

Existing deep interactive colorization models have focused on ways to utilize various types of interactions, such as point-wise color hints, scribbles, or natural-language texts, as methods to reflect a user's intent at runtime. However, another approach, which actively informs the user of the most effective regions to give hints for sketch image colorization, has been under-explored. This paper proposes a novel model-guided deep interactive colorization framework that reduces the required amount of user interactions, by prioritizing the regions in a colorization model. Our method, called GuidingPainter, prioritizes these regions where the model most needs a color hint, rather than just relying on the user's manual decision on where to give a color hint. In our extensive experiments, we show that our approach outperforms existing interactive colorization methods in terms of the conventional metrics, such as PSNR and FID, and reduces required amount of interactions.

Juntae Kim, Sung Min Ban

The spoof speech detection (SSD) is the essential countermeasure for automatic speaker verification systems. Although SSD with magnitude features in the frequency domain has shown promising results, the phase information also can be important to capture the artefacts of certain types of spoofing attacks. Thus, both magnitude and phase features must be considered to ensure the generalization ability to diverse types of spoofing attacks. In this paper, we investigate the failure reason of feature-level fusion of the previous works through the entropy analysis from which we found that the randomness difference between magnitude and phase features is large, which can interrupt the feature-level fusion via backend neural network; thus, we propose a phase network to reduce that difference. Our SSD system: phase network equipped Res2Net achieved significant performance improvement, specifically in the spoofing attack for which the phase information is considered to be important. Also, we demonstrate our SSD system in both known- and unknown-kind SSD scenarios for practical applications.

Juntae Kim, Minkyu Lim, Seokjin Hong

Inverse text normalization (ITN) is crucial for converting spoken-form into written-form, especially in the context of automatic speech recognition (ASR). While most downstream tasks of ASR rely on written-form, ASR systems often output spoken-form, highlighting the necessity for robust ITN in product-level ASR-based applications. Although neural ITN methods have shown promise, they still encounter performance challenges, particularly when dealing with ASR-generated spoken text. These challenges arise from the out-of-domain problem between training data and ASR-generated text. To address this, we propose a direct training approach that utilizes ASR-generated written or spoken text, with pairs augmented through ASR linguistic context emulation and a semi-supervised learning method enhanced by a large language model, respectively. Additionally, we introduce a post-aligning method to manage unpredictable errors, thereby enhancing the reliability of ITN. Our experiments show that our proposed methods remarkably improved ITN performance in various ASR scenarios.

Juntae Kim, Jeehye Lee

Recurrent neural network transducer (RNN-T) is an end-to-end speech recognition framework converting input acoustic frames into a character sequence. The state-of-the-art encoder network for RNN-T is the Conformer, which can effectively model the local-global context information via its convolution and self-attention layers. Although Conformer RNN-T has shown outstanding performance, most studies have been verified in the setting where the train and test data are drawn from the same domain. The domain mismatch problem for Conformer RNN-T has not been intensively investigated yet, which is an important issue for the product-level speech recognition system. In this study, we identified that fully connected self-attention layers in the Conformer caused high deletion errors, specifically in the long-form out-domain utterances. To address this problem, we introduce sparse self-attention layers for Conformer-based encoder networks, which can exploit local and generalized global information by pruning most of the in-domain fitted global connections. Also, we propose a state reset method for the generalization of the prediction network to cope with long-form utterances. Applying proposed methods to an out-domain test, we obtained 27.6% relative character error rate (CER) reduction compared to the fully connected self-attention layer-based Conformers.

Won Jang, Dan Lim, Jaesam Yoon et al.

Most neural vocoders employ band-limited mel-spectrograms to generate waveforms. If full-band spectral features are used as the input, the vocoder can be provided with as much acoustic information as possible. However, in some models employing full-band mel-spectrograms, an over-smoothing problem occurs as part of which non-sharp spectrograms are generated. To address this problem, we propose UnivNet, a neural vocoder that synthesizes high-fidelity waveforms in real time. Inspired by works in the field of voice activity detection, we added a multi-resolution spectrogram discriminator that employs multiple linear spectrogram magnitudes computed using various parameter sets. Using full-band mel-spectrograms as input, we expect to generate high-resolution signals by adding a discriminator that employs spectrograms of multiple resolutions as the input. In an evaluation on a dataset containing information on hundreds of speakers, UnivNet obtained the best objective and subjective results among competing models for both seen and unseen speakers. These results, including the best subjective score for text-to-speech, demonstrate the potential for fast adaptation to new speakers without a need for training from scratch.

Juntae Kim, Jaesung Bae, Minsoo Hahn

A state transition model (STM) based on chunk-wise classification was proposed for end-point detection (EPD). In general, EPD is developed using frame-wise voice activity detection (VAD) with additional STM, in which the state transition is conducted based on VAD's frame-level decision (speech or non-speech). However, VAD errors frequently occur in noisy environments, even though we use state-of-the-art deep neural network based VAD, which causes the undesired state transition of STM. In this work, to build robust STM, a state transition is conducted based on chunk-wise classification as EPD does not need to be conducted in frame-level. The chunk consists of multiple frames and the classification of chunk between speech and non-speech is done by aggregating the decisions of VAD for multiple frames, so that some undesired VAD errors in a chunk can be smoothed by other correct VAD decisions. Finally, the model was evaluated in both qualitative and quantitative measures including phone error rate.