René Heinrich

René Heinrich, Christoph Scholz, Stephan Vogt et al.

In recent years, researchers proposed a variety of deep learning models for wind power forecasting. These models predict the wind power generation of wind farms or entire regions more accurately than traditional machine learning algorithms or physical models. However, latest research has shown that deep learning models can often be manipulated by adversarial attacks. Since wind power forecasts are essential for the stability of modern power systems, it is important to protect them from this threat. In this work, we investigate the vulnerability of two different forecasting models to targeted, semi-targeted, and untargeted adversarial attacks. We consider a Long Short-Term Memory (LSTM) network for predicting the power generation of individual wind farms and a Convolutional Neural Network (CNN) for forecasting the wind power generation throughout Germany. Moreover, we propose the Total Adversarial Robustness Score (TARS), an evaluation metric for quantifying the robustness of regression models to targeted and semi-targeted adversarial attacks. It assesses the impact of attacks on the model's performance, as well as the extent to which the attacker's goal was achieved, by assigning a score between 0 (very vulnerable) and 1 (very robust). In our experiments, the LSTM forecasting model was fairly robust and achieved a TARS value of over 0.78 for all adversarial attacks investigated. The CNN forecasting model only achieved TARS values below 0.10 when trained ordinarily, and was thus very vulnerable. Yet, its robustness could be significantly improved by adversarial training, which always resulted in a TARS above 0.46.

Raphael Schwinger, Ben McEwen, Vincent S. Kather et al.

Passive acoustic monitoring enables large-scale biodiversity assessment, but reliable classification of bioacoustic sounds requires not only high accuracy but also well-calibrated uncertainty estimates to ground decision-making. In bioacoustics, calibration is challenged by overlapping vocalisations, long-tailed species distributions, and distribution shifts between training and deployment data. The calibration of multi-label deep learning classifiers within the domain of bioacoustics has not yet been assessed. We systematically benchmark the calibration of four state-of-the-art multi-label bird sound classifiers on the BirdSet benchmark, evaluating both global, per-dataset and per-class calibration using threshold-free calibration metrics (ECE, MCS) alongside discrimination metrics (cmAP). Model calibration varies significantly across datasets and classes. While Perch v2 and ConvNeXt$_{BS}$ show better global calibration, results vary between datasets. Both models indicate consistent underconfidence, while AudioProtoPNet and BirdMAE are mostly overconfident. Surprisingly, calibration seems to be better for less frequent classes. Using simple post hoc calibration methods we demonstrate a straightforward way to improve calibration. A small labelled calibration set is sufficient to significantly improve calibration with Platt scaling, while global calibration parameters suffer from dataset variability. Our findings highlight the importance of evaluating and improving uncertainty calibration in bioacoustic classifiers.

René Heinrich, Lukas Rauch, Bernhard Sick et al.

Deep learning models have significantly advanced acoustic bird monitoring by being able to recognize numerous bird species based on their vocalizations. However, traditional deep learning models are black boxes that provide no insight into their underlying computations, limiting their usefulness to ornithologists and machine learning engineers. Explainable models could facilitate debugging, knowledge discovery, trust, and interdisciplinary collaboration. This study introduces AudioProtoPNet, an adaptation of the Prototypical Part Network (ProtoPNet) for multi-label bird sound classification. It is an inherently interpretable model that uses a ConvNeXt backbone to extract embeddings, with the classification layer replaced by a prototype learning classifier trained on these embeddings. The classifier learns prototypical patterns of each bird species' vocalizations from spectrograms of training instances. During inference, audio recordings are classified by comparing them to the learned prototypes in the embedding space, providing explanations for the model's decisions and insights into the most informative embeddings of each bird species. The model was trained on the BirdSet training dataset, which consists of 9,734 bird species and over 6,800 hours of recordings. Its performance was evaluated on the seven test datasets of BirdSet, covering different geographical regions. AudioProtoPNet outperformed the state-of-the-art model Perch, achieving an average AUROC of 0.90 and a cmAP of 0.42, with relative improvements of 7.1% and 16.7% over Perch, respectively. These results demonstrate that even for the challenging task of multi-label bird sound classification, it is possible to develop powerful yet inherently interpretable deep learning models that provide valuable insights for ornithologists and machine learning engineers.

Lukas Rauch, Raphael Schwinger, Moritz Wirth et al.

Deep learning (DL) has greatly advanced audio classification, yet the field is limited by the scarcity of large-scale benchmark datasets that have propelled progress in other domains. While AudioSet is a pivotal step to bridge this gap as a universal-domain dataset, its restricted accessibility and limited range of evaluation use cases challenge its role as the sole resource. Therefore, we introduce BirdSet, a large-scale benchmark dataset for audio classification focusing on avian bioacoustics. BirdSet surpasses AudioSet with over 6,800 recording hours ($\uparrow\!17\%$) from nearly 10,000 classes ($\uparrow\!18\times$) for training and more than 400 hours ($\uparrow\!7\times$) across eight strongly labeled evaluation datasets. It serves as a versatile resource for use cases such as multi-label classification, covariate shift or self-supervised learning. We benchmark six well-known DL models in multi-label classification across three distinct training scenarios and outline further evaluation use cases in audio classification. We host our dataset on Hugging Face for easy accessibility and offer an extensive codebase to reproduce our results.

Lukas Rauch, René Heinrich, Ilyass Moummad et al.

Masked Autoencoders (MAEs) learn rich semantic representations in audio classification through an efficient self-supervised reconstruction task. However, general-purpose models fail to generalize well when applied directly to fine-grained audio domains. Specifically, bird-sound classification requires distinguishing subtle inter-species differences and managing high intra-species acoustic variability, revealing the performance limitations of general-domain Audio-MAEs. This work demonstrates that bridging this domain gap domain gap requires full-pipeline adaptation, not just domain-specific pretraining data. We systematically revisit and adapt the pretraining recipe, fine-tuning methods, and frozen feature utilization to bird sounds using BirdSet, a large-scale bioacoustic dataset comparable to AudioSet. Our resulting Bird-MAE achieves new state-of-the-art results in BirdSet's multi-label classification benchmark. Additionally, we introduce the parameter-efficient prototypical probing, enhancing the utility of frozen MAE representations and closely approaching fine-tuning performance in low-resource settings. Bird-MAE's prototypical probes outperform linear probing by up to 37 percentage points in mean average precision and narrow the gap to fine-tuning across BirdSet downstream tasks. Bird-MAE also demonstrates robust few-shot capabilities with prototypical probing in our newly established few-shot benchmark on BirdSet, highlighting the potential of tailored self-supervised learning pipelines for fine-grained audio domains.

Lukas Rauch, René Heinrich, Houtan Ghaffari et al.

Although probing frozen models has become a standard evaluation paradigm, self-supervised learning in audio defaults to fine-tuning. A key reason is that global pooling creates an information bottleneck causing linear probes to misrepresent the embedding quality: The $\texttt{cls}$-token discards crucial token information about dispersed, localized events in multi-label audio. This weakness is rooted in the mismatch between the pretraining objective (operating globally) and the downstream task (localized events). Across a comprehensive benchmark of 13 datasets and 6 spectrogram-based encoders, we first investigate the global pooling bottleneck. We then introduce binarized prototypical probes: a lightweight and simple pooling method that learns prototypes to perform class-wise information aggregation. Despite its simplicity, our method notably outperforms linear and attentive probing. Our work establishes probing as a competitive and efficient paradigm for evaluating audio SSL models, challenging the reliance on costly fine-tuning.

René Heinrich, Lukas Rauch, Bernhard Sick et al.

Adversarial training is a promising strategy for enhancing model robustness against adversarial attacks. However, its impact on generalization under substantial data distribution shifts in audio classification remains largely unexplored. To address this gap, this work investigates how different adversarial training strategies improve generalization performance and adversarial robustness in audio classification. The study focuses on two model architectures: a conventional convolutional neural network (ConvNeXt) and an inherently interpretable prototype-based model (AudioProtoPNet). The approach is evaluated using a challenging bird sound classification benchmark. This benchmark is characterized by pronounced distribution shifts between training and test data due to varying environmental conditions and recording methods, a common real-world challenge. The investigation explores two adversarial training strategies: one based on output-space attacks that maximize the classification loss function, and another based on embedding-space attacks designed to maximize embedding dissimilarity. These attack types are also used for robustness evaluation. Additionally, for AudioProtoPNet, the study assesses the stability of its learned prototypes under targeted embedding-space attacks. Results show that adversarial training, particularly using output-space attacks, improves clean test data performance by an average of 10.5% relative and simultaneously strengthens the adversarial robustness of the models. These findings, although derived from the bird sound domain, suggest that adversarial training holds potential to enhance robustness against both strong distribution shifts and adversarial attacks in challenging audio classification settings.

Malte Lehna, Björn Hoppmann, René Heinrich et al.

With the rising extension of renewable energies, the intraday electricity markets have recorded a growing popularity amongst traders as well as electric utilities to cope with the induced volatility of the energy supply. Through their short trading horizon and continuous nature, the intraday markets offer the ability to adjust trading decisions from the day-ahead market or reduce trading risk in a short-term notice. Producers of renewable energies utilize the intraday market to lower their forecast risk, by modifying their provided capacities based on current forecasts. However, the market dynamics are complex due to the fact that the power grids have to remain stable and electricity is only partly storable. Consequently, robust and intelligent trading strategies are required that are capable to operate in the intraday market. In this work, we propose a novel autonomous trading approach based on Deep Reinforcement Learning (DRL) algorithms as a possible solution. For this purpose, we model the intraday trade as a Markov Decision Problem (MDP) and employ the Proximal Policy Optimization (PPO) algorithm as our DRL approach. A simulation framework is introduced that enables the trading of the continuous intraday price in a resolution of one minute steps. We test our framework in a case study from the perspective of a wind park operator. We include next to general trade information both price and wind forecasts. On a test scenario of German intraday trading results from 2018, we are able to outperform multiple baselines with at least 45.24% improvement, showing the advantage of the DRL algorithm. However, we also discuss limitations and enhancements of the DRL agent, in order to increase the performance in future works.