Yulan Liu

SIMA team, Adrian Bolton, Alexander Lerchner et al.

We introduce SIMA 2, a generalist embodied agent that understands and acts in a wide variety of 3D virtual worlds. Built upon a Gemini foundation model, SIMA 2 represents a significant step toward active, goal-directed interaction within an embodied environment. Unlike prior work (e.g., SIMA 1) limited to simple language commands, SIMA 2 acts as an interactive partner, capable of reasoning about high-level goals, conversing with the user, and handling complex instructions given through language and images. Across a diverse portfolio of games, SIMA 2 substantially closes the gap with human performance and demonstrates robust generalization to previously unseen environments, all while retaining the base model's core reasoning capabilities. Furthermore, we demonstrate a capacity for open-ended self-improvement: by leveraging Gemini to generate tasks and provide rewards, SIMA 2 can autonomously learn new skills from scratch in a new environment. This work validates a path toward creating versatile and continuously learning agents for both virtual and, eventually, physical worlds.

Ilya Sklyar, Anna Piunova, Xianrui Zheng et al.

Automatic speech recognition (ASR) of single channel far-field recordings with an unknown number of speakers is traditionally tackled by cascaded modules. Recent research shows that end-to-end (E2E) multi-speaker ASR models can achieve superior recognition accuracy compared to modular systems. However, these models do not ensure real-time applicability due to their dependency on full audio context. This work takes real-time applicability as the first priority in model design and addresses a few challenges in previous work on multi-speaker recurrent neural network transducer (MS-RNN-T). First, we introduce on-the-fly overlapping speech simulation during training, yielding 14% relative word error rate (WER) improvement on LibriSpeechMix test set. Second, we propose a novel multi-turn RNN-T (MT-RNN-T) model with an overlap-based target arrangement strategy that generalizes to an arbitrary number of speakers without changes in the model architecture. We investigate the impact of the maximum number of speakers seen during training on MT-RNN-T performance on LibriCSS test set, and report 28% relative WER improvement over the two-speaker MS-RNN-T. Third, we experiment with a rich transcription strategy for joint recognition and segmentation of multi-party speech. Through an in-depth analysis, we discuss potential pitfalls of the proposed system as well as promising future research directions.

Amin Fazel, Wei Yang, Yulan Liu et al.

End-to-end (E2E) automatic speech recognition (ASR) models have recently demonstrated superior performance over the traditional hybrid ASR models. Training an E2E ASR model requires a large amount of data which is not only expensive but may also raise dependency on production data. At the same time, synthetic speech generated by the state-of-the-art text-to-speech (TTS) engines has advanced to near-human naturalness. In this work, we propose to utilize synthetic speech for ASR training (SynthASR) in applications where data is sparse or hard to get for ASR model training. In addition, we apply continual learning with a novel multi-stage training strategy to address catastrophic forgetting, achieved by a mix of weighted multi-style training, data augmentation, encoder freezing, and parameter regularization. In our experiments conducted on in-house datasets for a new application of recognizing medication names, training ASR RNN-T models with synthetic audio via the proposed multi-stage training improved the recognition performance on new application by more than 65% relative, without degradation on existing general applications. Our observations show that SynthASR holds great promise in training the state-of-the-art large-scale E2E ASR models for new applications while reducing the costs and dependency on production data.

Manuel Giollo, Deniz Gunceler, Yulan Liu et al.

Bootstrapping speech recognition on limited data resources has been an area of active research for long. The recent transition to all-neural models and end-to-end (E2E) training brought along particular challenges as these models are known to be data hungry, but also came with opportunities around language-agnostic representations derived from multilingual data as well as shared word-piece output representations across languages that share script and roots. We investigate here the effectiveness of different strategies to bootstrap an RNN-Transducer (RNN-T) based automatic speech recognition (ASR) system in the low resource regime, while exploiting the abundant resources available in other languages as well as the synthetic audio from a text-to-speech (TTS) engine. Our experiments demonstrate that transfer learning from a multilingual model, using a post-ASR text-to-text mapping and synthetic audio deliver additive improvements, allowing us to bootstrap a model for a new language with a fraction of the data that would otherwise be needed. The best system achieved a 46% relative word error rate (WER) reduction compared to the monolingual baseline, among which 25% relative WER improvement is attributed to the post-ASR text-to-text mappings and the TTS synthetic data.

Ilya Sklyar, Anna Piunova, Yulan Liu

Recent research shows end-to-end ASR systems can recognize overlapped speech from multiple speakers. However, all published works have assumed no latency constraints during inference, which does not hold for most voice assistant interactions. This work focuses on multi-speaker speech recognition based on a recurrent neural network transducer (RNN-T) that has been shown to provide high recognition accuracy at a low latency online recognition regime. We investigate two approaches to multi-speaker model training of the RNN-T: deterministic output-target assignment and permutation invariant training. We show that guiding separation with speaker order labels in the former case enhances the high-level speaker tracking capability of RNN-T. Apart from that, with multistyle training on single- and multi-speaker utterances, the resulting models gain robustness against ambiguous numbers of speakers during inference. Our best model achieves a WER of 10.2% on simulated 2-speaker LibriSpeech data, which is competitive with the previously reported state-of-the-art nonstreaming model (10.3%), while the proposed model could be directly applied for streaming applications.

Xianrui Zheng, Yulan Liu, Deniz Gunceler et al.

Today, many state-of-the-art automatic speech recognition (ASR) systems apply all-neural models that map audio to word sequences trained end-to-end along one global optimisation criterion in a fully data driven fashion. These models allow high precision ASR for domains and words represented in the training material but have difficulties recognising words that are rarely or not at all represented during training, i.e. trending words and new named entities. In this paper, we use a text-to-speech (TTS) engine to provide synthetic audio for out-of-vocabulary (OOV) words. We aim to boost the recognition accuracy of a recurrent neural network transducer (RNN-T) on OOV words by using the extra audio-text pairs, while maintaining the performance on the non-OOV words. Different regularisation techniques are explored and the best performance is achieved by fine-tuning the RNN-T on both original training data and extra synthetic data with elastic weight consolidation (EWC) applied on the encoder. This yields a 57% relative word error rate (WER) reduction on utterances containing OOV words without any degradation on the whole test set.