Google DeepMind has introduced a powerful on-device version of its Gemini Robotics AI. This new system allows robots to complete complex tasks without relying on a cloud connection. Known as Gemini Robotics On-Device, the model brings Gemini's advanced reasoning and control capabilities directly into physical robots. It is designed for fast, reliable performance in places with poor or no internet connectivity, making it ideal for real-world, latency-sensitive environments. Unlike its cloud-connected predecessor, this version runs entirely on the robot itself.

When we think about what makes an audiobook memorable, it's always the most human moments: a catch in the throat when tears are near, or words spoken through a real smile. A Melbourne actor and audiobook narrator, Annabelle Tudor, says it's the instinct we have as storytellers that makes narration such a primal, and precious, skill. "The voice betrays how we're feeling really easily," she says. But as an art form it may be under threat. In May the Amazon-owned audiobook provider Audible announced it would allow authors and publishers to choose from more than 100 voices created by artificial intelligence to narrate audiobooks in English, Spanish, French and Italian, with AI translation of audiobooks expected to be available later in the year – news that was met with criticism and curiosity across the publishing industry.

Developing a single foundation model with the capability to excel across diverse tasks has been a long-standing objective in the field of artificial intelligence. As the wave of general-purpose foundation models sweeps across various domains, their influence has significantly extended to the field of recommendation systems. While recent efforts have explored recommendation foundation models for various generative tasks, they often overlook crucial embedding tasks and struggle with the complexities of multi-task learning, including knowledge sharing & conflict resolution, and convergence speed inconsistencies. To address these limitations, we introduce RecFound, a generative representational learning framework for recommendation foundation models. We construct the first comprehensive dataset for recommendation foundation models covering both generative and embedding tasks across diverse scenarios. Based on this dataset, we propose a novel multi-task training scheme featuring a Task-wise Mixture of Low-rank Experts (TMoLE) to handle knowledge sharing & conflict, a Step-wise Convergence-oriented Sample Scheduler (S2Sched) to address inconsistent convergence, and a Model Merge module to balance the performance across tasks. Experiments demonstrate that RecFound achieves state-of-the-art performance across various recommendation tasks, outperforming existing baselines.

We introduce Nirantar, a comprehensive framework for evaluating continual learning (CL) in multilingual and multi-domain ASR. Designed to reflect real-world CL challenges, Nirantar leverages data collected incrementally across 22 languages and 208 districts in India through natural episodes. This enables evaluation across Language-Incremental (LIL), Domain-Incremental (DIL), and the novel Language-Incremental Domain-Incremental Learning (LIDIL) scenarios. Unlike prior work that relies on simulated episodes, Nirantar presents dynamic, non-uniform language and domain shifts, making it an ideal testbed for CL research. With 3250 hours of human-transcribed speech, including 1720 hours newly introduced in this work, our framework enables systematic benchmarking of CL methods. We evaluate existing approaches and demonstrate that no single method performs consistently well, underscoring the need for more robust CL strategies.

The idea that Gregorian melodies are constructed from some vocabulary of segments has long been a part of chant scholarship. This so-called "centonisation" theory has received much musicological criticism, but frequent re-use of certain melodic segments has been observed in chant melodies, and the intractable number of possible segmentations allowed the option that some undiscovered segmentation exists that will yet prove the value of centonisation, and recent empirical results have shown that segmentations can outperform music-theoretical features in mode classification. Inspired by the fact that Gregorian chant was memorised, we search for an optimal unsupervised segmentation of chant melody using nested hierarchical Pitman-Yor language models. The segmentation we find achieves state-of-the-art performance in mode classification. Modeling a monk memorising the melodies from one liturgical manuscript, we then find empirical evidence for the link between mode classification and memory efficiency, and observe more formulaic areas at the beginnings and ends of melodies corresponding to the practical role of modality in performance. However, the resulting segmentations themselves indicate that even such a memory-optimal segmentation is not what is understood as centonisation.

The visual world around us constantly evolves, from real-time news and social media trends to global infrastructure changes visible through satellite imagery and augmented reality enhancements. However, Multimodal Large Language Models (MLLMs), which automate many tasks, struggle to stay current, limited by the cutoff dates in their fixed training datasets. To quantify this stagnation, we introduce LiveVQA, the first-of-its-kind dataset featuring 107,143 samples and 12 categories data specifically designed to support research in both seeking and updating with live visual knowledge. Drawing from recent news articles, video platforms, and academic publications in April 2024-May 2025, LiveVQA enables evaluation of how models handle latest visual information beyond their knowledge boundaries and how current methods help to update them. Our comprehensive benchmarking of 17 state-of-the-art MLLMs reveals significant performance gaps on content beyond knowledge cutoff, and tool-use or agentic visual seeking framework drastically gain an average of 327% improvement. Furthermore, we explore parameter-efficient fine-tuning (PEFT) methods to update MLLMs with new visual knowledge. We dive deeply to the critical balance between adapter capacity and model capability when updating MLLMs with new visual knowledge. All the experimental dataset and source code are publicly available at: https://livevqa.github.io.

More than twenty-five years ago, first ideas were developed on how to design a system that can provide recommendations to groups of users instead of individual users. Since then, a rich variety of algorithmic proposals were published, e.g., on how to acquire individual preferences, how to aggregate them, and how to generate recommendations for groups of users. However, despite the rich literature on the topic, barely any examples of real-world group recommender systems can be found. This lets us question common assumptions in academic research, in particular regarding communication processes in a group and how recommendation-supported decisions are made. In this essay, we argue that these common assumptions and corresponding system designs often may not match the needs or expectations of users. We thus call for a reorientation in this research area, leveraging the capabilities of modern Generative AI assistants like ChatGPT. Specifically, as one promising future direction, we envision group recommender systems to be systems where human group members interact in a chat and an AI-based group recommendation agent assists the decision-making process in an agentic way. Ultimately, this shall lead to a more natural group decision-making environment and finally to wider adoption of group recommendation systems in practice.

This paper introduces von Mises-Fisher exploration (vMF-exp), a scalable method for exploring large action sets in reinforcement learning problems where hyperspherical embedding vectors represent these actions. vMF-exp involves initially sampling a state embedding representation using a von Mises-Fisher distribution, then exploring this representation's nearest neighbors, which scales to virtually unlimited numbers of candidate actions. We show that, under theoretical assumptions, vMF-exp asymptotically maintains the same probability of exploring each action as Boltzmann Exploration (B-exp), a popular alternative that, nonetheless, suffers from scalability issues as it requires computing softmax values for each action. Consequently, vMF-exp serves as a scalable alternative to B-exp for exploring large action sets with hyperspherical embeddings. Experiments on simulated data, real-world public data, and the successful large-scale deployment of vMF-exp on the recommender system of a global music streaming service empirically validate the key properties of the proposed method.

Beat tracking in musical performance MIDI is a challenging and important task for notation-level music transcription and rhythmical analysis, yet existing methods primarily focus on audio-based approaches. This paper proposes an end-to-end transformer-based model for beat and downbeat tracking in performance MIDI, leveraging an encoder-decoder architecture for sequence-to-sequence translation of MIDI input to beat annotations. Our approach introduces novel data preprocessing techniques, including dynamic augmentation and optimized tokenization strategies, to improve accuracy and generalizability across different datasets. We conduct extensive experiments using the A-MAPS, ASAP, GuitarSet, and Leduc datasets, comparing our model against state-of-the-art hidden Markov models (HMMs) and deep learning-based beat tracking methods. The results demonstrate that our model outperforms existing symbolic music beat tracking approaches, achieving competitive F1-scores across various musical styles and instruments. Our findings highlight the potential of transformer architectures for symbolic beat tracking and suggest future integration with automatic music transcription systems for enhanced music analysis and score generation.

Recent advances in word embeddings and language models use large-scale, unlabelled data and self-supervised learning to boost NLP performance. Multilingual models, often trained on web-sourced data like Wikipedia, face challenges: few low-resource languages are included, their data is often noisy, and lack of labeled datasets makes it hard to evaluate performance outside high-resource languages like English. In this dissertation, we focus on languages spoken in Sub-Saharan Africa where all the indigenous languages in this region can be regarded as low-resourced in terms of the availability of labelled data for NLP tasks and unlabelled data found on the web. We analyse the noise in the publicly available corpora, and curate a high-quality corpus, demonstrating that the quality of semantic representations learned in word embeddings does not only depend on the amount of data but on the quality of pre-training data. We demonstrate empirically the limitations of word embeddings, and the opportunities the multilingual pre-trained language model (PLM) offers especially for languages unseen during pre-training and low-resource scenarios. We further study how to adapt and specialize multilingual PLMs to unseen African languages using a small amount of monolingual texts. To address the under-representation of the African languages in NLP research, we developed large scale human-annotated labelled datasets for 21 African languages in two impactful NLP tasks: named entity recognition and machine translation. We conduct an extensive empirical evaluation using state-of-the-art methods across supervised, weakly-supervised, and transfer learning settings.