Advancements in artificial intelligence (AI) have significantly enhanced the realism and interactivity of non-player characters (NPCs) in virtual reality (VR), creating more engaging and believable user experiences. This paper evaluates AI-driven NPCs within a VR interrogation simulator, focusing on their perceived realism, usability, and system performance. The simulator features two AI-powered NPCs, a suspect, and a partner, using GPT-4 Turbo to engage participants in a scenario to determine the suspect's guilt or innocence. A user study with 18 participants assessed the system using the System Usability Scale (SUS), Game Experience Questionnaire (GEQ), and a Virtual Agent Believability Questionnaire, alongside latency measurements for speech-to-text (STT), text-to-speech (TTS), OpenAI GPT-4 Turbo, and overall (cycle) latency. Results showed an average cycle latency of 7 seconds, influenced by the increasing conversational context. Believability scored 6.67 out of 10, with high ratings in behavior, social relationships, and intelligence but moderate scores in emotion and personality. The system achieved a SUS score of 79.44, indicating good usability. These findings demonstrate the potential of large language models to improve NPC realism and interaction in VR while highlighting challenges in reducing system latency and enhancing emotional depth. This research contributes to the development of more sophisticated AI-driven NPCs, revealing the need for performance optimization to achieve increasingly immersive virtual experiences.
The scarcity of high-quality annotated data presents a significant challenge in developing effective end-to-end speech-to-text translation (ST) systems, particularly for low-resource languages. This paper explores the hypothesis that weakly labeled data can be used to build ST models for low-resource language pairs. We constructed speech-to-text translation datasets with the help of bitext mining using state-of-the-art sentence encoders. We mined the multilingual Shrutilipi corpus to build Shrutilipi-anuvaad, a dataset comprising ST data for language pairs Bengali-Hindi, Malayalam-Hindi, Odia-Hindi, and Telugu-Hindi. We created multiple versions of training data with varying degrees of quality and quantity to investigate the effect of quality versus quantity of weakly labeled data on ST model performance. Results demonstrate that ST systems can be built using weakly labeled data, with performance comparable to massive multi-modal multilingual baselines such as SONAR and SeamlessM4T.
With the rapid progress of speech language models (SLMs), discrete speech tokens have emerged as a core interface between speech and text, enabling unified modeling across modalities. Recent speech tokenization approaches aim to isolate semantic information from low-level acoustics to better align with language models. In particular, previous methods use SSL teachers such as HuBERT to extract semantic representations, which are then distilled into a semantic quantizer to suppress acoustic redundancy as well as capture content-related latent structures. However, they still produce speech token sequences significantly longer than their textual counterparts, creating challenges for efficient speech-language modeling. Reducing the frame rate is a natural solution, but standard techniques, such as rigid average pooling across frames, can distort or dilute the semantic structure required for effective LM alignment. To address this, we propose LM-SPT, a speech tokenization method that introduces a novel semantic distillation. Instead of directly matching teacher and student features via pooling, we reconstruct speech solely from semantic tokens and minimize the discrepancy between the encoded representations of the original and reconstructed waveforms, obtained from a frozen automatic speech recognition (ASR) encoder. This indirect yet data-driven supervision enables the tokenizer to learn discrete units that are more semantically aligned with language models. LM-SPT further incorporates architectural improvements to the encoder and decoder for speech tokenization, and supports multiple frame rates, including 25Hz, 12.5Hz, and 6.25Hz. Experimental results show that LM-SPT achieves superior reconstruction fidelity compared to baselines, and that SLMs trained with LM-SPT tokens achieve competitive performances on speech-to-text and consistently outperform baselines on text-to-speech tasks.
Existing 3D visual grounding methods rely on precise text prompts to locate objects within 3D scenes. Speech, as a natural and intuitive modality, offers a promising alternative. Real-world speech inputs, however, often suffer from transcription errors due to accents, background noise, and varying speech rates, limiting the applicability of existing 3DVG methods. To address these challenges, we propose \textbf{SpeechRefer}, a novel 3DVG framework designed to enhance performance in the presence of noisy and ambiguous speech-to-text transcriptions. SpeechRefer integrates seamlessly with xisting 3DVG models and introduces two key innovations. First, the Speech Complementary Module captures acoustic similarities between phonetically related words and highlights subtle distinctions, generating complementary proposal scores from the speech signal. This reduces dependence on potentially erroneous transcriptions. Second, the Contrastive Complementary Module employs contrastive learning to align erroneous text features with corresponding speech features, ensuring robust performance even when transcription errors dominate. Extensive experiments on the SpeechRefer and peechNr3D datasets demonstrate that SpeechRefer improves the performance of existing 3DVG methods by a large margin, which highlights SpeechRefer's potential to bridge the gap between noisy speech inputs and reliable 3DVG, enabling more intuitive and practical multimodal systems.
Multilingual speech-to-speech translation (S2ST) aims to directly convert spoken utterances from multiple source languages into fluent and intelligible speech in a target language. Despite recent progress, several critical challenges persist: 1) achieving high-quality S2ST remains a significant obstacle; 2) most existing S2ST methods rely heavily on large-scale parallel speech corpora, which are difficult and resource-intensive to obtain. To tackle these challenges, we introduce S2ST-Omni, a novel, efficient, and scalable framework tailored for multilingual speech-to-speech translation. Specifically, we decompose S2ST into speech-to-text translation (S2TT) and text-to-speech synthesis (TTS). To enable high-quality S2TT while mitigating reliance on large-scale parallel speech corpora, we leverage powerful pretrained models: Whisper for robust audio understanding and Qwen 3.0 for advanced text comprehension. A lightweight speech adapter is introduced to bridge the modality gap between speech and text representations, facilitating effective utilization of pretrained multimodal knowledge. To ensure both translation accuracy and real-time responsiveness, we adopt a streaming speech generation model in the TTS stage, which generates the target speech in an autoregressive manner. Extensive experiments conducted on the CVSS benchmark demonstrate that S2ST-Omni consistently surpasses several state-of-the-art S2ST baselines in translation quality, highlighting its effectiveness and superiority.
Swiss German is a low-resource language represented by diverse dialects that differ significantly from Standard German and from each other, lacking a standardized written form. As a result, transcribing Swiss German involves translating into Standard German. Existing datasets have been collected in controlled environments, yielding effective speech-to-text (STT) models, but these models struggle with spontaneous conversational speech. This paper, therefore, introduces the new SRB-300 dataset, a 300-hour annotated speech corpus featuring real-world long-audio recordings from 39 Swiss German radio and TV stations. It captures spontaneous speech across all major Swiss dialects recorded in various realistic environments and overcomes the limitation of prior sentence-level corpora. We fine-tuned multiple OpenAI Whisper models on the SRB-300 dataset, achieving notable enhancements over previous zero-shot performance metrics. Improvements in word error rate (WER) ranged from 19% to 33%, while BLEU scores increased between 8% and 40%. The best fine-tuned model, large-v3, achieved a WER of 17.1% and a BLEU score of 74.8. This advancement is crucial for developing effective and robust STT systems for Swiss German and other low-resource languages in real-world contexts.
Current translation systems, despite being highly multilingual, cover only 5% of the world's languages. Expanding language coverage to the long-tail of low-resource languages requires data-efficient methods that rely on cross-lingual and cross-modal knowledge transfer. To this end, we propose a character-based approach to improve adaptability to new languages and modalities. Our method leverages SONAR, a multilingual fixed-size embedding space with different modules for encoding and decoding. We use a teacher-student approach with parallel translation data to obtain a character-level encoder. Then, using ASR data, we train a lightweight adapter to connect a massively multilingual CTC ASR model (MMS), to the character-level encoder, potentially enabling speech translation from 1,000+ languages. Experimental results in text translation for 75 languages on FLORES+ demonstrate that our character-based approach can achieve better language transfer than traditional subword-based models, especially outperforming them in low-resource settings, and demonstrating better zero-shot generalizability to unseen languages. Our speech adaptation, maximizing knowledge transfer from the text modality, achieves state-of-the-art results in speech-to-text translation on the FLEURS benchmark on 33 languages, surpassing previous supervised and cascade models, albeit being a zero-shot model with minimal supervision from ASR data.
This paper discusses the construction, fine-tuning, and deployment of BeaverTalk, a cascaded system for speech-to-text translation as part of the IWSLT 2025 simultaneous translation task. The system architecture employs a VAD segmenter for breaking a speech stream into segments, Whisper Large V2 for automatic speech recognition (ASR), and Gemma 3 12B for simultaneous translation. Regarding the simultaneous translation LLM, it is fine-tuned via low-rank adaptors (LoRAs) for a conversational prompting strategy that leverages a single prior-sentence memory bank from the source language as context. The cascaded system participated in the English$\rightarrow$German and English$\rightarrow$Chinese language directions for both the low and high latency regimes. In particular, on the English$\rightarrow$German task, the system achieves a BLEU of 24.64 and 27.83 at a StreamLAAL of 1837.86 and 3343.73, respectively. Then, on the English$\rightarrow$Chinese task, the system achieves a BLEU of 34.07 and 37.23 at a StreamLAAL of 2216.99 and 3521.35, respectively.
We propose a Speech-to-Text Translation (S2TT) approach that integrates phoneme representations into a Chain-of-Thought (CoT) framework to improve translation in low-resource and zero-resource settings. By introducing phoneme recognition as an intermediate step, we enhance cross-lingual transfer, enabling translation even for languages with no labeled speech data. Our system builds on a multilingual LLM, which we extend to process speech and phonemes. Training follows a curriculum learning strategy that progressively introduces more complex tasks. Experiments on multilingual S2TT benchmarks show that phoneme-augmented CoT improves translation quality in low-resource conditions and enables zero-resource translation, while slightly impacting high-resource performance. Despite this trade-off, our findings demonstrate that phoneme-based CoT is a promising step toward making S2TT more accessible across diverse languages.
Training large-scale models presents challenges not only in terms of resource requirements but also in terms of their convergence. For this reason, the learning rate (LR) is often decreased when the size of a model is increased. Such a simple solution is not enough in the case of speech-to-text (S2T) trainings, where evolved and more complex variants of the Transformer architecture -- e.g., Conformer or Branchformer -- are used in light of their better performance. As a workaround, OWSM designed a double linear warmup of the LR, increasing it to a very small value in the first phase before updating it to a higher value in the second phase. While this solution worked well in practice, it was not compared with alternative solutions, nor was the impact on the final performance of different LR warmup schedules studied. This paper fills this gap, revealing that i) large-scale S2T trainings demand a sub-exponential LR warmup, and ii) a higher LR in the warmup phase accelerates initial convergence, but it does not boost final performance.
Recent advances in Multimodal Large Language Models (MLLMs) have significantly enhanced the naturalness and flexibility of human computer interaction by enabling seamless understanding across text, vision, and audio modalities. Among these, voice enabled models such as SpeechGPT have demonstrated considerable improvements in usability, offering expressive, and emotionally responsive interactions that foster deeper connections in real world communication scenarios. However, the use of voice introduces new security risks, as attackers can exploit the unique characteristics of spoken language, such as timing, pronunciation variability, and speech to text translation, to craft inputs that bypass defenses in ways not seen in text-based systems. Despite substantial research on text based jailbreaks, the voice modality remains largely underexplored in terms of both attack strategies and defense mechanisms. In this work, we present an adversarial attack targeting the speech input of aligned MLLMs in a white box scenario. Specifically, we introduce a novel token level attack that leverages access to the model's speech tokenization to generate adversarial token sequences. These sequences are then synthesized into audio prompts, which effectively bypass alignment safeguards and to induce prohibited outputs. Evaluated on SpeechGPT, our approach achieves up to 89 percent attack success rate across multiple restricted tasks, significantly outperforming existing voice based jailbreak methods. Our findings shed light on the vulnerabilities of voice-enabled multimodal systems and to help guide the development of more robust next-generation MLLMs.
This paper introduces a novel method for automated server provisioning by integrating Transformerbased Named Entity Recognition models with Automated Speech Detection using OpenAI's Whisper. Leveraging advanced Transformer architectures-BERT, RoBERTa, and DeBERTa-combined with robust speech-to-text capabilities, our approach enables IT professionals to provision cloud servers efficiently via natural spoken commands. A customannotated dataset containing real-world and AI-generated provisioning requests is presented, meticulously labeled using the BIO tagging scheme across fourteen critical entity categories relevant to cloud infrastructure provisioning. Comprehensive evaluations of model performance and robustness were conducted under realistic conditions, including controlled transcription noise to simulate practical speech recognition errors. While all tested models achieved high performance on clean test data, results from noisy test scenarios revealed notable disparities in model generalization capabilities. Specifically, DeBERTa exhibited exceptional resilience, maintaining an F1 score of 96.23 % under adverse conditions. These findings highlight the practicality and robustness of combining speech-to-text processing with advanced Named Entity Recognition models, significantly advancing real-time, voice-driven IT automation workflows.
In today's digitally-driven world, the demand for personalized and context-aware recommendations has never been greater. Traditional recommender systems have made significant strides in this direction, but they often lack the ability to tap into the richness of conversational data. This paper represents a novel approach to recommendation systems by integrating conversational insights into the recommendation process. The Conversational Recommender System integrates cutting-edge technologies such as deep learning, leveraging machine learning algorithms like Apriori for Association Rule Mining, Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Long Short-Term Memory (LTSM). Furthermore, sophisticated voice recognition technologies, including Hidden Markov Models (HMMs) and Dynamic Time Warping (DTW) algorithms, play a crucial role in accurate speech-to-text conversion, ensuring robust performance in diverse environments. The methodology incorporates a fusion of content-based and collaborative recommendation approaches, enhancing them with NLP techniques. This innovative integration ensures a more personalized and context-aware recommendation experience, particularly in marketing applications.
In recent years, end-to-end speech-to-speech (S2S) dialogue systems have garnered increasing research attention due to their advantages over traditional cascaded systems, including achieving lower latency and more natural integration of nonverbal cues such as emotion and speaker identity. However, these end-to-end systems face key challenges, particularly in incorporating external knowledge, a capability commonly addressed by Retrieval-Augmented Generation (RAG) in text-based large language models (LLMs). The core difficulty lies in the modality gap between input speech and retrieved textual knowledge, which hinders effective integration. To address this issue, we propose a novel end-to-end RAG framework that directly retrieves relevant textual knowledge from speech queries, eliminating the need for intermediate speech-to-text conversion via techniques like ASR. Experimental results demonstrate that our method significantly improves the performance of end-to-end S2S dialogue systems while achieving higher retrieval efficiency. Although the overall performance still lags behind cascaded models, our framework offers a promising direction for enhancing knowledge integration in end-to-end S2S systems. We will release the code and dataset to support reproducibility and promote further research in this area.
Accurate intrinsic and extrinsic camera calibration can be an important prerequisite for robotic applications that rely on vision as input. While there is ongoing research on enabling camera calibration using natural images, many systems in practice still rely on using designated calibration targets with e.g. checkerboard patterns or April tag grids. Once calibration images from different perspectives have been acquired and feature descriptors detected, those are typically used in an optimization process to minimize the geometric reprojection error. For this optimization to converge, input images need to be of sufficient quality and particularly sharpness; they should neither contain motion blur nor rolling-shutter artifacts that can arise when the calibration board was not static during image capture. In this work, we present a novel calibration image acquisition technique controlled via voice commands recorded with a clip-on microphone, that can be more robust and user-friendly than e.g. triggering capture with a remote control, or filtering out blurry frames from a video sequence in postprocessing. To achieve this, we use a state-of-the-art speech-to-text transcription model with accurate per-word timestamping to capture trigger words with precise temporal alignment. Our experiments show that the proposed method improves user experience by being fast and efficient, allowing us to successfully calibrate complex multi-camera setups.
Developing Automatic Speech Recognition (ASR) systems for Tunisian Arabic Dialect is challenging due to the dialect's linguistic complexity and the scarcity of annotated speech datasets. To address these challenges, we propose the LinTO audio and textual datasets -- comprehensive resources that capture phonological and lexical features of Tunisian Arabic Dialect. These datasets include a variety of texts from numerous sources and real-world audio samples featuring diverse speakers and code-switching between Tunisian Arabic Dialect and English or French. By providing high-quality audio paired with precise transcriptions, the LinTO audio and textual datasets aim to provide qualitative material to build and benchmark ASR systems for the Tunisian Arabic Dialect. Keywords -- Tunisian Arabic Dialect, Speech-to-Text, Low-Resource Languages, Audio Data Augmentation
Although speech recognition algorithms have developed quickly in recent years, achieving high transcription accuracy across diverse audio formats and acoustic environments remains a major challenge. This work explores how incorporating custom language models with the open-source Vosk Toolkit can improve speech-to-text accuracy in varied settings. Unlike many conventional systems limited to specific audio types, this approach supports multiple audio formats such as WAV, MP3, FLAC, and OGG by using Python modules for preprocessing and format conversion. A Python-based transcription pipeline was developed to process input audio, perform speech recognition using Vosk's KaldiRecognizer, and export the output to a DOCX file. Results showed that custom models reduced word error rates, especially in domain-specific scenarios involving technical terminology, varied accents, or background noise. This work presents a cost-effective, offline solution for high-accuracy transcription and opens up future opportunities for automation and real-time applications.
Open-Ended Learning (OEL) autonomous robots can acquire new skills and knowledge through direct interaction with their environment, relying on mechanisms such as intrinsic motivations and self-generated goals to guide learning processes. OEL robots are highly relevant for applications as they can autonomously leverage acquired knowledge to perform tasks beneficial to human users in unstructured environments, addressing challenges unforeseen at design time. However, OEL robots face a significant limitation: their openness may lead them to waste time learning information that is irrelevant to tasks desired by specific users. Here, we propose a solution called `Purpose-Directed Open-Ended Learning' (POEL), based on the novel concept of `purpose' introduced in previous work. A purpose specifies what users want the robot to achieve. The key insight of this work is that purpose can focus OEL on learning self-generated classes of tasks that, while unknown during autonomous learning (as typical in OEL), involve objects relevant to the purpose. This concept is operationalised in a novel robot architecture capable of receiving a human purpose through speech-to-text, analysing the scene to identify objects, and using a Large Language Model to reason about which objects are purpose-relevant. These objects are then used to bias OEL exploration towards their spatial proximity and to self-generate rewards that favour interactions with them. The solution is tested in a simulated scenario where a camera-arm-gripper robot interacts freely with purpose-related and distractor objects. For the first time, the results demonstrate the potential advantages of purpose-focused OEL over state-of-the-art OEL methods, enabling robots to handle unstructured environments while steering their learning toward knowledge acquisition relevant to users.
In end-to-end speech translation, acoustic representations learned by the encoder are usually fixed and static, from the perspective of the decoder, which is not desirable for dealing with the cross-modal and cross-lingual challenge in speech translation. In this paper, we show the benefits of varying acoustic states according to decoder hidden states and propose an adaptive speech-to-text translation model that is able to dynamically adapt acoustic states in the decoder. We concatenate the acoustic state and target word embedding sequence and feed the concatenated sequence into subsequent blocks in the decoder. In order to model the deep interaction between acoustic states and target hidden states, a speech-text mixed attention sublayer is introduced to replace the conventional cross-attention network. Experiment results on two widely-used datasets show that the proposed method significantly outperforms state-of-the-art neural speech translation models.
This work proposes an industry-level omni-modal large language model (LLM) pipeline that integrates auditory, visual, and linguistic modalities to overcome challenges such as limited tri-modal datasets, high computational costs, and complex feature alignments. Our pipeline consists of three main components: First, a modular framework enabling flexible configuration of various encoder-LLM-decoder architectures. Second, a lightweight training strategy that pre-trains audio-language alignment on the state-of-the-art vision-language model Qwen2.5-VL, thus avoiding the costly pre-training of vision-specific modalities. Third, an audio synthesis pipeline that generates high-quality audio-text data from diverse real-world scenarios, supporting applications such as Automatic Speech Recognition and Speech-to-Speech chat. To this end, we introduce an industry-level omni-modal LLM, Nexus. Extensive experiments validate the efficacy of our pipeline, yielding the following key findings:(1) In the visual understanding task, Nexus exhibits superior performance compared with its backbone model - Qwen2.5-VL-7B, validating the efficiency of our training strategy. (2) Within the English Spoken Question-Answering task, the model achieves better accuracy than the same-period competitor (i.e, MiniCPM-o2.6-7B) in the LLaMA Q. benchmark. (3) In our real-world ASR testset, Nexus achieves outstanding performance, indicating its robustness in real scenarios. (4) In the Speech-to-Text Translation task, our model outperforms Qwen2-Audio-Instruct-7B. (5) In the Text-to-Speech task, based on pretrained vocoder (e.g., Fishspeech1.4 or CosyVoice2.0), Nexus is comparable to its backbone vocoder on Seed-TTS benchmark. (6) An in-depth analysis of tri-modal alignment reveals that incorporating the audio modality enhances representational alignment between vision and language.