Recent literature uses language to build foundation models for audio. These Audio-Language Models (ALMs) are trained on a vast number of audio-text pairs and show remarkable performance in tasks including Text-to-Audio Retrieval, Captioning, and Question Answering. However, their ability to engage in more complex open-ended tasks, like Interactive Question-Answering, requires proficiency in logical reasoning- a skill not yet benchmarked. We introduce the novel task of Audio Entailment to evaluate an ALM's deductive reasoning ability. This task assesses whether a text description (hypothesis) of audio content can be deduced from an audio recording (premise), with potential conclusions being entailment, neutral, or contradiction, depending on the sufficiency of the evidence. We create two datasets for this task with audio recordings sourced from two audio captioning datasets-AudioCaps and Clotho-and hypotheses generated using Large Language Models (LLMs). We benchmark state-of-the-art ALMs and find deficiencies in logical reasoning with both zero-shot and linear probe evaluations. Finally, we propose "caption-before-reason", an intermediate step of captioning that improves the Zero-Shot and linear-probe performance of ALMs by an absolute 6% and 3%, respectively.

Automated Audio Captioning (AAC) is the task of generating natural language descriptions given an audio stream. A typical AAC system requires manually curated training data of audio segments and corresponding text caption annotations. The creation of these audio-caption pairs is costly, resulting in general data scarcity for the task. In this work, we address this major limitation and propose an approach to train AAC systems using only text. Our approach leverages the multimodal space of contrastively trained audio-text models, such as CLAP. During training, a decoder generates captions conditioned on the pretrained CLAP text encoder. During inference, the text encoder is replaced with the pretrained CLAP audio encoder. To bridge the modality gap between text and audio embeddings, we propose the use of noise injection or a learnable adapter, during training. We find that the proposed text-only framework performs competitively with stateof-the-art models trained with paired audio, showing that efficient text-to-audio transfer is possible. Finally, we showcase both stylized audio captioning and caption enrichment while training without audio or human-created text captions.

Emotions lie on a continuum, but current models treat emotions as a finite valued discrete variable. This representation does not capture the diversity in the expression of emotion. To better represent emotions we propose the use of natural language descriptions (or prompts). In this work, we address the challenge of automatically generating these prompts and training a model to better learn emotion representations from audio and prompt pairs. We use acoustic properties that are correlated to emotion like pitch, intensity, speech rate, and articulation rate to automatically generate prompts i.e. ‘acoustic prompts’. We use a contrastive learning objective to map speech to their respective acoustic prompts. We evaluate our model on Emotion Audio Retrieval and Speech Emotion Recognition. Our results show that the acoustic prompts significantly improve the model’s performance in EAR, in various Precision@K metrics. In SER, we observe a 3.8% relative accuracy improvement on the Ravdess dataset.

Audio-Language models jointly learn multimodal text and audio representations that enable Zero-Shot inference. Models rely on the encoders to create powerful representations of the input and generalize to multiple tasks ranging from sounds, music, and speech. Although models have achieved remarkable performance, there is still a gap with task-specific models. In this paper, we propose a Contrastive Language-Audio Pretraining model that is pretrained with a diverse collection of 4.6M audio-text pairs employing two innovative encoders for Zero-Shot inference. To learn audio representations, we trained an audio encoder on 22 audio tasks, instead of the standard training of sound event classification. To learn language representations, we trained an autoregressive decoder-only model instead of the standard encoder-only models. Then, the audio and language representations are brought into a joint multimodal space using Contrastive Learning. We used our encoders to improve the downstream performance by a large margin. We extensively evaluated the generalization of our representations on 26 downstream tasks, the largest in the literature. Our model achieves state of the art results in several tasks outperforming 4 different models and leading the way towards general-purpose audio representations. Code is on GitHub1.

Mainstream machine listening models are trained to learn audio concepts under the paradigm of one class label to many recordings focusing on one task. Learning under such restricted supervision limits the flexibility of models because they require labeled audio for training and can only predict the predefined categories. Instead, we propose to learn audio concepts from natural language supervision. We call our approach Contrastive Language-Audio Pretraining (CLAP), which connects language and audio by using two encoders and a contrastive learning objective, bringing audio and text descriptions into a joint multimodal space. We trained CLAP with 128k audio and text pairs and evaluated it on 16 downstream tasks across 7 domains, such as classification of sound events, scenes, music, and speech. CLAP establishes state-of-the-art (SoTA) in Zero-Shot performance. Also, we evaluated CLAP’s audio encoder in a supervised learning setup and achieved SoTA in 5 tasks. The Zero-Shot capability removes the need of training with class labeled audio, enables flexible class prediction at inference time, and generalizes well in multiple downstream tasks. Code is available at: https://github.com/microsoft/CLAP.

Psychological research has postulated that emotions and sentiment are correlated to dimensional scores of valence, arousal, and dominance. However, the literature of Speech Emotion Recognition focuses on independently predicting the three of them for a given speech audio. In this paper, we evaluate and quantify the predictive power of the dimensional scores towards categorical emotions and sentiment for two publicly available speech emotion datasets. We utilize the three emotional views in a joined multi-view training framework. The views comprise the dimensional scores, emotions categories, and sentiment categories. We present a comparison for each emotional view or combination of, utilizing two general-purpose models for speech-related applications: CNN14 and Wav2Vec2. To our knowledge this is the first time such a joint framework is explored. We found that a joined multi-view training framework can produce results as strong or stronger than models trained independently for each view.

In the domain of audio processing, Transfer Learning has facilitated the rise of Self-Supervised Learning and Zero-Shot Learning techniques. These approaches have led to the development of versatile models capable of tackling a wide array of tasks, while delivering state-of-the-art performance. However, current models inherently lack the capacity to produce the requisite language for open-ended tasks, such as Audio Captioning or Audio Question Answering. We introduce Pengi, a novel Audio Language Model that leverages Transfer Learning by framing all audio tasks as text-generation tasks. It takes as input, an audio recording, and text, and generates free-form text as output. The input audio is represented as a sequence of continuous embeddings by an audio encoder. A text encoder does the same for the corresponding text input. Both sequences are combined as a prefix to prompt a pre-trained frozen language model. The unified architecture of Pengi enables open-ended tasks and close-ended tasks without any additional fine-tuning or task-specific extensions. When evaluated on 21 downstream tasks, our approach yields state-of-the-art performance in several of them. Our results show that connecting language models with audio models is a major step towards general-purpose audio understanding.

Realistic recordings of soundscapes often have multiple sound events co-occurring, such as car horns, engine and human voices. Sound event retrieval is a type of contentbased search aiming at finding audio samples, similar to an audio query based on their acoustic or semantic content. State of the art sound event retrieval models have focused on single-label audio recordings, with only one sound event occurring, rather than on multi-label audio recordings (i.e., multiple sound events occur in one recording). To address this latter problem, we propose different Deep Learning architectures with a Siamesestructure and a Pairwise Presence Matrix. The networks are trained and evaluated using the SONYC-UST dataset containing both single- and multi-label soundscape recordings. The performance results show the effectiveness of our proposed model.

Existing audio search engines use one of two approaches: matching text-text or audio-audio pairs. In the former, text queries are matched to semantically similar words in an index of audio metadata to retrieve corresponding audio clips or segments, while in the latter, audio signals are directly used to retrieve acoustically-similar recordings from an audio database. However, independent treatment of text and audio has precluded information exchange between the two modalities. This is a problem because similarity in language does not always imply similarity in acoustics, and vice versa. Moreover, independent modeling can be error prone especially for ad hoc, user-generated recordings, which are noisy in both audio and their associated textual labels. To overcome this limitation, we propose a framework that learns joint embeddings from a shared lexico-acoustic space, where vectors from either modality can be mapped together and compared directly. Thus, we improve semantic knowledge and enable the use of either text or audio queries to search and retrieve audio. Our results break new ground for a cross-modal audio search engine, and further exploration of lexico-acoustic spaces.

Each edition of the challenge on Detection and Classification of Acoustic Scenes and Events (DCASE) contained several tasks involving sound event detection in different setups. DCASE 2017 presented participants with three such tasks, each having specific datasets and detection requirements: Task 2, in which target sound events were very rare in both training and testing data, Task 3 having overlapping events annotated in real-life audio, and Task 4, in which only weakly labeled data were available for training. In this paper, we present three tasks, including the datasets and baseline systems, and analyze the challenge entries for each task. We observe the popularity of methods using deep neural networks, and the still widely used mel frequency-based representations, with only few approaches standing out as radically different. Analysis of the systems behavior reveals that task-specific optimization has a big role in producing good performance; however, often this optimization closely follows the ranking metric, and its maximization/minimization does not result in universally good performance. We also introduce the calculation of confidence intervals based on a jackknife resampling procedure to perform statistical analysis of the challenge results. The analysis indicates that while the 95% confidence intervals for many systems overlap, there are significant differences in performance between the top systems and the baseline for all tasks.

The largest source of sound events is web videos. Most videos lack sound event labels at segment level, however, a significant number of them do respond to text queries, from a match found using metadata by search engines. In this paper we explore the extent to which a search query can be used as the true label for detection of sound events in videos. We present a framework for large-scale sound event recognition on web videos. The framework crawls videos using search queries corresponding to 78 sound event labels drawn from three datasets. The datasets are used to train three classifiers, and we obtain a prediction on 3.7 million web video segments. We evaluated performance using the search query as true label and compare it with human labeling. Both types of ground truth exhibited close performance, to within 10%, and similar performance trend with increasing number of evaluated segments. Hence, our experiments show potential for using search query as a preliminary true label for sound event recognition in web videos.

State of the art acoustic scene classification techniques often employ features of large dimensionality, which are then used to train and perform inferences with kernel machines such as Support Vector Machines. However, the complexity of computing the non-linear kernel matrix for these methods increases with the dimensionality of the features and the size of the dataset. In this work, we introduce a new scheme that hashes features, which combined with a linear function approximates a non-linear Laplacian kernel. Each hash typically has lower dimensionality than the input features and each component is represented by one bit instead of floating values. Hence, allowing efficient computation of the kernel matrix using XOR operations rather than dot-products. Our scheme is demonstrated mathematically and tested in the 2017 DCASE: Acoustic Scene Classification. The hashes reduce up to six powers of two the feature representation with minimal loss of accuracy.

In this paper, we focus on the problem of content-based retrieval for audio, which aims to retrieve all semantically similar audio recordings for a given audio clip query. This problem is similar to the problem of query by example of audio, which aims to retrieve media samples from a database, which are similar to the user-provided example. We propose a novel approach which encodes the audio into a vector representation using Siamese Neural Networks. The goal is to obtain an encoding similar for files belonging to the same audio class, thus allowing retrieval of semantically similar audio. Using simple similarity measures such as those based on simple euclidean distance and cosine similarity we show that these representations can be very effectively used for retrieving recordings similar in audio content.

In this paper we propose methods to extract geographically relevant information in a multimedia recording using its audio. Our method primarily is based on the fact that urban acoustic environment consists of a variety of sounds. Hence, location information can be inferred from the composition of sound events/classes present in the audio. More specifically, we adopt matrix factorization techniques to obtain semantic content of recording in terms of different sound classes. These semantic information are then combined to identify the location of recording.

This paper presents advances in analyzing audio content information to detect events in videos, such as a parade or a birthday party. We developed a set of tools for audio processing within the predominantly vision-focused deep neural network (DNN) framework Caffe. Using these tools, we show, for the first time, the potential of using only a DNN for audio-based multimedia event detection. Training DNNs for event detection using the entire audio track from each video causes a computational bottleneck. Here, we address this problem by developing a sparse audio frame-sampling method that improves event-detection speed and accuracy. We achieved a 10 percentage-point improvement in event classification accuracy, with a 200x reduction in the number of training input examples as compared to using the entire track. This reduction in input feature volume led to a 16x reduction in the size of the DNN architecture and a 300x reduction in training time. We applied our method using the recently released YLI-MED dataset and compared our results with a state-of-the-art system and with results reported in the literature for TRECVIDMED. Our results show much higher MAP scores compared to a baseline i-vector system - at a significantly reduced computational cost. The speed improvement is relevant for processing videos on a large scale, and could enable more effective deployment in mobile systems.

Traditional speech/non-speech segmentation systems have been designed for specific acoustic conditions, such as broadcast news or meetings. However, little research has been done on consumer-produced audio. This type of media is constantly growing and has complex characteristics such as low quality recordings, environmental noise and overlapping sounds. This paper discusses an evaluation of three different approaches for speech/non-speech detection on consumer-produced audio. The approaches are state-of-the-art speech/non-speech detectors-one based on Gaussian Mixture Models (GMM), another on Support Vector Machines (SVM), and the last on Neural Networks (NN). Using the TRECVID MED 2012 database, we designed training/testing sets combinations to aid the understanding of what speech/non-speech detection on consumer-produced media entails and how traditional approaches to this detection performed in this domain. The results revealed that the cross-domain state-of-the-art GMM and SVM systems' tests underperformed a one-layer NN algorithm, which had 20% higher accuracy and computed audio 5 times faster.

Audio-based video event detection (VED) on user-generated content (UGC) aims to find videos that show an observable event such as a wedding ceremony or birthday party rather than a sound, such as music, clapping or singing. The difficulty of video content analysis on UGC lies in the acoustic variability and lack of structure of the data. The UGC task has been explored mainly by computer vision, but can be benefited by the used of audio. The i-vector system is state-of-the-art in Speaker Verification, and is outperforming a conventional Gaussian Mixture Model (GMM)-based approach. The system compensates for undesired acoustic variability and extracts information from the acoustic environment, making it a meaningful choice for detection on UGC. This paper employs the i-vector-based system for audio-based VED on UGC and expands the understanding of the system on the task. It also includes a performance comparison with the conventional GMM-based and state-of-the-art Random Forest (RF)-based systems. The i-vector system aids audio-based event detection by addressing UGC audio characteristics. It outperforms the GMM-based system, and is competitive with the RF-based system in terms of the Missed Detection (MD) rate at 4% and 2.8% False Alarm (FA) rates, and complements the RF-based system by demonstrating slightly improvement in combination over the standalone systems.