arXiv subjects

Markus Schedl

Publications and source records attributed to Markus Schedl.

At least 37 records · Page 2Linked to original sources

arXiv2026

How Far Can Pretrained LLMs Go in Symbolic Music? Controlled Comparisons of Supervised and Preference-based Adaptation

Music often shares notable parallels with language, motivating the use of pretrained large language models (LLMs) for symbolic music understanding and generation. Despite growing interest, the practical effectiveness of adapting instruction-tuned LLMs to symbolic music remains insufficiently characterized. We present a controlled comparative study of finetuning strategies for ABC-based generation and understanding, comparing an off-the-shelf instruction-tuned backbone to domain-adapted variants and a music-specialized LLM baseline. Across multiple symbolic music corpora and evaluation signals, we provide some insights into adaptation choices for symbolic music applications. We highlight the domain adaptation vs.~preserving prior information tradeoff as well as the distinct behaviour of metrics used to measure the domain adaptation for symbolic music.

cs.SD↗

arXiv2026

Distillation-based Layer Dropping (DLD): Effective End-to-end Framework for Dynamic Speech Networks

Edge devices operate in constrained and varying resource settings, requiring dynamic architectures that can adapt to limitations of the available resources. To meet such demands, layer dropping ($\mathcal{LD}$) approach is typically used to transform static models into dynamic ones by skipping parts of the network along with reducing overall computational complexity. However, existing $\mathcal{LD}$ methods greatly impact the dynamic model's performance for low and high dropping cases, deteriorating the performance-computation trade-off. To this end, we propose a distillation-based layer dropping (DLD) framework that effectively combines the capabilities of knowledge distillation and $\mathcal{LD}$ in an end-to-end fashion, thereby achieving state-of-the-art performance for dynamic speech networks. Comprehensive experimentation utilizing well-known speech recognition methods, including conformer and WavLM, on three public benchmarks demonstrates the effectiveness of our framework, reducing the word error rate by $9.32\%$ and $2.25\%$ for high and no dropping cases with $33.3\%$ reduction in training time.

cs.SD↗

arXiv2026

Towards Fair ASR For Second Language Speakers Using Fairness Prompted Finetuning

In this work, we address the challenge of building fair English ASR systems for second-language speakers. Our analysis of widely used ASR models, Whisper and Seamless-M4T, reveals large fluctuations in word error rate (WER) across 26 accent groups, indicating significant fairness gaps. To mitigate this, we propose fairness-prompted finetuning with lightweight adapters, incorporating Spectral Decoupling (SD), Group Distributionally Robust Optimization (Group-DRO), and Invariant Risk Minimization (IRM). Our proposed fusion of traditional empirical risk minimization (ERM) with cross-entropy and fairness-driven objectives (SD, Group DRO, and IRM) enhances fairness across accent groups while maintaining overall recognition accuracy. In terms of macro-averaged word error rate, our approach achieves a relative improvement of 58.7% and 58.5% over the large pretrained Whisper and SeamlessM4T, and 9.7% and 7.8% over them, finetuning with standard empirical risk minimization with cross-entropy loss.

cs.CL↗

arXiv2026

Face-Voice Association with Inductive Bias for Maximum Class Separation

Face-voice association is widely studied in multimodal learning and is approached representing faces and voices with embeddings that are close for a same person and well separated from those of others. Previous work achieved this with loss functions. Recent advancements in classification have shown that the discriminative ability of embeddings can be strengthened by imposing maximum class separation as inductive bias. This technique has never been used in the domain of face-voice association, and this work aims at filling this gap. More specifically, we develop a method for face-voice association that imposes maximum class separation among multimodal representations of different speakers as an inductive bias. Through quantitative experiments we demonstrate the effectiveness of our approach, showing that it achieves SOTA performance on two task formulation of face-voice association. Furthermore, we carry out an ablation study to show that imposing inductive bias is most effective when combined with losses for inter-class orthogonality. To the best of our knowledge, this work is the first that applies and demonstrates the effectiveness of maximum class separation as an inductive bias in multimodal learning; it hence paves the way to establish a new paradigm.

cs.CV↗

arXiv2025

Linking Faces and Voices Across Languages: Insights from the FAME 2026 Challenge

Over half of the world's population is bilingual and people often communicate under multilingual scenarios. The Face-Voice Association in Multilingual Environments (FAME) 2026 Challenge, held at ICASSP 2026, focuses on developing methods for face-voice association that are effective when the language at test-time is different than the training one. This report provides a brief summary of the challenge.

cs.CV↗

arXiv2025

RobustA: Robust Anomaly Detection in Multimodal Data

In recent years, multimodal anomaly detection methods have demonstrated remarkable performance improvements over video-only models. However, real-world multimodal data is often corrupted due to unforeseen environmental distortions. In this paper, we present the first-of-its-kind work that comprehensively investigates the adverse effects of corrupted modalities on multimodal anomaly detection task. To streamline this work, we propose RobustA, a carefully curated evaluation dataset to systematically observe the impacts of audio and visual corruptions on the overall effectiveness of anomaly detection systems. Furthermore, we propose a multimodal anomaly detection method, which shows notable resilience against corrupted modalities. The proposed method learns a shared representation space for different modalities and employs a dynamic weighting scheme during inference based on the estimated level of corruption. Our work represents a significant step forward in enabling the real-world application of multimodal anomaly detection, addressing situations where the likely events of modality corruptions occur. The proposed evaluation dataset with corrupted modalities and respective extracted features will be made publicly available.

cs.LG↗

arXiv2025

Face-voice Association in Multilingual Environments (FAME) 2026 Challenge Evaluation Plan

The advancements of technology have led to the use of multimodal systems in various real-world applications. Among them, audio-visual systems are among the most widely used multimodal systems. In the recent years, associating face and voice of a person has gained attention due to the presence of unique correlation between them. The Face-voice Association in Multilingual Environments (FAME) 2026 Challenge focuses on exploring face-voice association under the unique condition of a multilingual scenario. This condition is inspired from the fact that half of the world's population is bilingual and most often people communicate under multilingual scenarios. The challenge uses a dataset named Multilingual Audio-Visual (MAV-Celeb) for exploring face-voice association in multilingual environments. This report provides the details of the challenge, dataset, baseline models, and task details for the FAME Challenge.

cs.CV↗

arXiv2025

Single-Branch Network Architectures to Close the Modality Gap in Multimodal Recommendation

Traditional recommender systems rely on collaborative filtering, using past user-item interactions to help users discover new items in a vast collection. In cold start, i.e., when interaction histories of users or items are not available, content-based recommender systems use side information instead. Hybrid recommender systems (HRSs) often employ multimodal learning to combine collaborative and side information, which we jointly refer to as modalities. Though HRSs can provide recommendations when some modalities are missing, their quality degrades. In this work, we utilize single-branch neural networks equipped with weight sharing, modality sampling, and contrastive loss to provide accurate recommendations even in missing modality scenarios by narrowing the modality gap. We compare these networks with multi-branch alternatives and conduct extensive experiments on three datasets. Six accuracy-based and four beyond-accuracy-based metrics help assess the recommendation quality for the different training paradigms and their hyperparameters in warm-start and missing modality scenarios. We quantitatively and qualitatively study the effects of these different aspects on bridging the modality gap. Our results show that single-branch networks achieve competitive performance in warm-start scenarios and are significantly better in missing modality settings. Moreover, our approach leads to closer proximity of an item's modalities in the embedding space. Our full experimental setup is available at https://github.com/hcai-mms/single-branch-networks.

cs.IR↗

arXiv2025

Music4All A+A: A Multimodal Dataset for Music Information Retrieval Tasks

Music is characterized by aspects related to different modalities, such as the audio signal, the lyrics, or the music video clips. This has motivated the development of multimodal datasets and methods for Music Information Retrieval (MIR) tasks such as genre classification or autotagging. Music can be described at different levels of granularity, for instance defining genres at the level of artists or music albums. However, most datasets for multimodal MIR neglect this aspect and provide data at the level of individual music tracks. We aim to fill this gap by providing Music4All Artist and Album (Music4All A+A), a dataset for multimodal MIR tasks based on music artists and albums. Music4All A+A is built on top of the Music4All-Onion dataset, an existing track-level dataset for MIR tasks. Music4All A+A provides metadata, genre labels, image representations, and textual descriptors for 6,741 artists and 19,511 albums. Furthermore, since Music4All A+A is built on top of Music4All-Onion, it allows access to other multimodal data at the track level, including user--item interaction data. This renders Music4All A+A suitable for a broad range of MIR tasks, including multimodal music recommendation, at several levels of granularity. To showcase the use of Music4All A+A, we carry out experiments on multimodal genre classification of artists and albums, including an analysis in missing-modality scenarios, and a quantitative comparison with genre classification in the movie domain. Our experiments show that images are more informative for classifying the genres of artists and albums, and that several multimodal models for genre classification struggle in generalizing across domains. We provide the code to reproduce our experiments at https://github.com/hcai-mms/Music4All-A-A, the dataset is linked in the repository and provided open-source under a CC BY-NC-SA 4.0 license.

cs.MM↗

arXiv2025

Investigating Gender Bias in LLM-Generated Stories via Psychological Stereotypes

As Large Language Models (LLMs) are increasingly used across different applications, concerns about their potential to amplify gender biases in various tasks are rising. Prior research has often probed gender bias using explicit gender cues as counterfactual, or studied them in sentence completion and short question answering tasks. These formats might overlook more implicit forms of bias embedded in generative behavior of longer content. In this work, we investigate gender bias in LLMs using gender stereotypes studied in psychology (e.g., aggressiveness or gossiping) in an open-ended task of narrative generation. We introduce a novel dataset called StereoBias-Stories containing short stories either unconditioned or conditioned on (one, two, or six) random attributes from 25 psychological stereotypes and three task-related story endings. We analyze how the gender contribution in the overall story changes in response to these attributes and present three key findings: (1) While models, on average, are highly biased towards male in unconditioned prompts, conditioning on attributes independent from gender stereotypes mitigates this bias. (2) Combining multiple attributes associated with the same gender stereotype intensifies model behavior, with male ones amplifying bias and female ones alleviating it. (3) Model biases align with psychological ground-truth used for categorization, and alignment strength increases with model size. Together, these insights highlight the importance of psychology-grounded evaluation of LLMs.

cs.CL↗

arXiv2025

Parameter-Efficient Single Collaborative Branch for Recommendation

Recommender Systems (RS) often rely on representations of users and items in a joint embedding space and on a similarity metric to compute relevance scores. In modern RS, the modules to obtain user and item representations consist of two distinct and separate neural networks (NN). In multimodal representation learning, weight sharing has been proven effective in reducing the distance between multiple modalities of a same item. Inspired by these approaches, we propose a novel RS that leverages weight sharing between the user and item NN modules used to obtain the latent representations in the shared embedding space. The proposed framework consists of a single Collaborative Branch for Recommendation (CoBraR). We evaluate CoBraR by means of quantitative experiments on e-commerce and movie recommendation. Our experiments show that by reducing the number of parameters and improving beyond-accuracy aspects without compromising accuracy, CoBraR has the potential to be applied and extended for real-world scenarios.

cs.IR↗

arXiv2025

Just Ask for Music (JAM): Multimodal and Personalized Natural Language Music Recommendation

Natural language interfaces offer a compelling approach for music recommendation, enabling users to express complex preferences conversationally. While Large Language Models (LLMs) show promise in this direction, their scalability in recommender systems is limited by high costs and latency. Retrieval-based approaches using smaller language models mitigate these issues but often rely on single-modal item representations, overlook long-term user preferences, and require full model retraining, posing challenges for real-world deployment. In this paper, we present JAM (Just Ask for Music), a lightweight and intuitive framework for natural language music recommendation. JAM models user-query-item interactions as vector translations in a shared latent space, inspired by knowledge graph embedding methods like TransE. To capture the complexity of music and user intent, JAM aggregates multimodal item features via cross-attention and sparse mixture-of-experts. We also introduce JAMSessions, a new dataset of over 100k user-query-item triples with anonymized user/item embeddings, uniquely combining conversational queries and user long-term preferences. Our results show that JAM provides accurate recommendations, produces intuitive representations suitable for practical use cases, and can be easily integrated with existing music recommendation stacks.

cs.IR↗

arXiv2025

Double Entendre: Robust Audio-Based AI-Generated Lyrics Detection via Multi-View Fusion

The rapid advancement of AI-based music generation tools is revolutionizing the music industry but also posing challenges to artists, copyright holders, and providers alike. This necessitates reliable methods for detecting such AI-generated content. However, existing detectors, relying on either audio or lyrics, face key practical limitations: audio-based detectors fail to generalize to new or unseen generators and are vulnerable to audio perturbations; lyrics-based methods require cleanly formatted and accurate lyrics, unavailable in practice. To overcome these limitations, we propose a novel, practically grounded approach: a multimodal, modular late-fusion pipeline that combines automatically transcribed sung lyrics and speech features capturing lyrics-related information within the audio. By relying on lyrical aspects directly from audio, our method enhances robustness, mitigates susceptibility to low-level artifacts, and enables practical applicability. Experiments show that our method, DE-detect, outperforms existing lyrics-based detectors while also being more robust to audio perturbations. Thus, it offers an effective, robust solution for detecting AI-generated music in real-world scenarios. Our code is available at https://github.com/deezer/robust-AI-lyrics-detection.

cs.CL↗

arXiv2025

AI-Generated Song Detection via Lyrics Transcripts

The recent rise in capabilities of AI-based music generation tools has created an upheaval in the music industry, necessitating the creation of accurate methods to detect such AI-generated content. This can be done using audio-based detectors; however, it has been shown that they struggle to generalize to unseen generators or when the audio is perturbed. Furthermore, recent work used accurate and cleanly formatted lyrics sourced from a lyrics provider database to detect AI-generated music. However, in practice, such perfect lyrics are not available (only the audio is); this leaves a substantial gap in applicability in real-life use cases. In this work, we instead propose solving this gap by transcribing songs using general automatic speech recognition (ASR) models. We do this using several detectors. The results on diverse, multi-genre, and multi-lingual lyrics show generally strong detection performance across languages and genres, particularly for our best-performing model using Whisper large-v2 and LLM2Vec embeddings. In addition, we show that our method is more robust than state-of-the-art audio-based ones when the audio is perturbed in different ways and when evaluated on different music generators. Our code is available at https://github.com/deezer/robust-AI-lyrics-detection.

cs.SD↗

arXiv2025

PAEFF: Precise Alignment and Enhanced Gated Feature Fusion for Face-Voice Association

We study the task of learning association between faces and voices, which is gaining interest in the multimodal community lately. These methods suffer from the deliberate crafting of negative mining procedures as well as the reliance on the distant margin parameter. These issues are addressed by learning a joint embedding space in which orthogonality constraints are applied to the fused embeddings of faces and voices. However, embedding spaces of faces and voices possess different characteristics and require spaces to be aligned before fusing them. To this end, we propose a method that accurately aligns the embedding spaces and fuses them with an enhanced gated fusion thereby improving the performance of face-voice association. Extensive experiments on the VoxCeleb dataset reveals the merits of the proposed approach.

cs.CV↗

arXiv2025

Hybrid Personalization Using Declarative and Procedural Memory Modules of the Cognitive Architecture ACT-R

Recommender systems often rely on sub-symbolic machine learning approaches that operate as opaque black boxes. These approaches typically fail to account for the cognitive processes that shape user preferences and decision-making. In this vision paper, we propose a hybrid user modeling framework based on the cognitive architecture ACT-R that integrates symbolic and sub-symbolic representations of human memory. Our goal is to combine ACT-R's declarative memory, which is responsible for storing symbolic chunks along sub-symbolic activations, with its procedural memory, which contains symbolic production rules. This integration will help simulate how users retrieve past experiences and apply decision-making strategies. With this approach, we aim to provide more transparent recommendations, enable rule-based explanations, and facilitate the modeling of cognitive biases. We argue that our approach has the potential to inform the design of a new generation of human-centered, psychology-informed recommender systems.

cs.IR↗

arXiv2025

Familiarizing with Music: Discovery Patterns for Different Music Discovery Needs

Humans have the tendency to discover and explore. This natural tendency is reflected in data from streaming platforms as the amount of previously unknown content accessed by users. Additionally, in domains such as that of music streaming there is evidence that recommending novel content improves users' experience with the platform. Therefore, understanding users' discovery patterns, such as the amount to which and the way users access previously unknown content, is a topic of relevance for both the scientific community and the streaming industry, particularly the music one. Previous works studied how music consumption differs for users of different traits and looked at diversity, novelty, and consistency over time of users' music preferences. However, very little is known about how users discover and explore previously unknown music, and how this behavior differs for users of varying discovery needs. In this paper we bridge this gap by analyzing data from a survey answered by users of the major music streaming platform Deezer in combination with their streaming data. We first address questions regarding whether users who declare a higher interest in unfamiliar music listen to more diverse music, have more stable music preferences over time, and explore more music within a same time window, compared to those who declare a lower interest. We then investigate which type of music tracks users choose to listen to when they explore unfamiliar music, identifying clear patterns of popularity and genre representativeness that vary for users of different discovery needs. Our findings open up possibilities to infer users' interest in unfamiliar music from streaming data as well as possibilities to develop recommender systems that guide users in exploring music in a more natural way.

cs.IR↗

arXiv2025

Unsupervised Graph Embeddings for Session-based Recommendation with Item Features

In session-based recommender systems, predictions are based on the user's preceding behavior in the session. State-of-the-art sequential recommendation algorithms either use graph neural networks to model sessions in a graph or leverage the similarity of sessions by exploiting item features. In this paper, we combine these two approaches and propose a novel method, Graph Convolutional Network Extension (GCNext), which incorporates item features directly into the graph representation via graph convolutional networks. GCNext creates a feature-rich item co-occurrence graph and learns the corresponding item embeddings in an unsupervised manner. We show on three datasets that integrating GCNext into sequential recommendation algorithms significantly boosts the performance of nearest-neighbor methods as well as neural network models. Our flexible extension is easy to incorporate in state-of-the-art methods and increases the MRR@20 by up to 12.79%.

cs.IR↗