Coding-based assessment is essential not only for evaluating students' programming skills but also for assessing their ability to apply theoretical concepts to real-world scenarios. Jupyter Notebooks constitute a unique platform to assess coding-based pieces of work. Using a Jupyter Notebook, students can write, execute and visualise code while documenting their pro-cess through explanations, equations, and multimedia elements. However, identifying cases of plagiarism and collusion in Jupyter Notebooks can be challenging. Since Jupyter Notebooks cannot be processed by traditional text-based plagiarism detection tools, the ability to identify plagiarism largely relies on manual approaches. Hence, to fully benefit from Jupyter Notebooks as an assessment platform, automatic approaches that can contribute to identifying plagiarism and collusion are needed. In this paper, we introduce a new tool - JBEval, that identifies similar code and text blocks in a collection of Jupyter Notebooks and provides similarity scores for overall code and text individually. JBEval also provides a visual comparison of any two Jupyter Notebooks, allowing to trace identical blocks for further investigation and reporting. Our similarity detection algorithm is parameterized for controlling the aggressiveness of the similarity-based detection process, which makes it adaptable to a wide diversity of courses. This fast and reliable tool will be made widely accessible by sharing it in the public domain, enabling educators across diverse disciplines to integrate it into their assessment workflows.

Conventional music structure analysis algorithms aim to divide a song into segments and to group them with abstract labels (e.g., ‘A’, ‘B’, and ‘C’). However, explicitly identifying the function of each segment (e.g., ‘verse’ or ‘chorus’) is rarely attempted, but has many applications. We introduce a multi-task deep learning framework to model these structural semantic labels directly from audio by estimating "verseness," "chorusness," and so forth, as a function of time. We propose a 7-class taxonomy (i.e., intro, verse, chorus, bridge, outro, instrumental, and silence) and provide rules to consolidate annotations from four disparate datasets. We also propose to use a spectral-temporal Transformer-based model, called SpecTNT, which can be trained with an additional connectionist temporal localization (CTL) loss. In cross-dataset evaluations using four public datasets, we demonstrate the effectiveness of the SpecTNT model and CTL loss, and obtain strong results overall: the proposed system outperforms state-of-the-art chorus-detection and boundary-detection methods at detecting choruses and boundaries, respectively.

A music mashup combines audio elements from two or more songs to create a new work. To reduce the time and effort required to make them, researchers have developed algorithms that predict the compatibility of audio elements. Prior work has focused on mixing unaltered excerpts, but advances in source separation enable the creation of mashups from isolated stems (e.g., vocals, drums, bass, etc.). In this work, we take advantage of separated stems not just for creating mashups, but for training a model that predicts the mutual compatibility of groups of excerpts, using self-supervised and semi-supervised methods. Specifically, we first produce a random mashup creation pipeline that combines stem tracks obtained via source separation, with key and tempo automatically adjusted to match, since these are prerequisites for high-quality mashups. To train a model to predict compatibility, we use stem tracks obtained from the same song as positive examples, and random combinations of stems with key and/or tempo unadjusted as negative examples. To improve the model and use more data, we also train on "average" examples: random combinations with matching key and tempo, where we treat them as unlabeled data as their true compatibility is unknown. To determine whether the combined signal or the set of stem signals is more indicative of the quality of the result, we experiment on two model architectures and train them using semi-supervised learning technique. Finally, we conduct objective and subjective evaluations of the system, comparing them to a standard rule-based system.

This paper presents a novel supervised approach to detecting the chorus segments in popular music. Traditional approaches to this task are mostly unsupervised, with pipelines designed to target some quality that is assumed to define "chorusness," which usually means seeking the loudest or most frequently repeated sections. We propose to use a convolutional neural network with a multi-task learning objective, which simultaneously fits two temporal activation curves: one indicating "chorusness" as a function of time, and the other the location of the boundaries. We also propose a post-processing method that jointly takes into account the chorus and boundary predictions to produce binary output. In experiments using three datasets, we compare our system to a set of public implementations of other segmentation and chorus-detection algorithms, and find our approach performs significantly better.

We propose a music structure analysis method that converts a path-enhanced self-similarity matrix (SSM) into a block-enhanced SSM using non-negative matrix factor 2-D deconvolution (NMF2D). With a non-negative constraint, the deconvolution intuitively corresponds to the repeated stripes in the path-enhanced SSM. Then the block-enhanced SSM is constructed without any clustering technique. We fuse block-enhanced SSMs obtained using different parameters, resulting in better and more robust results. Discussion shows that the proposed method can be a potential tool for analysing music structure at different scales.

The prevalence of exact repetition in loop-based music makes it an opportune target for source separation. Nonnegative factorization approaches have been used to model the repetition of looped content, and kernel additive modeling has leveraged periodicity within a piece to separate looped background elements. We propose a novel method of leveraging periodicity in a factorization model: we treat the two-dimensional spectrogram as a three-dimensional tensor, and use nonnegative tensor factorization to estimate the component spectral templates, rhythms and loop recurrences in a single step. Testing our method on synthesized loop-based examples, we find that our algorithm mostly exceeds the performance of competing methods, with a reduction in execution cost. We discuss limitations of the algorithm as we demonstrate its potential to analyze larger and more complex songs.

Transcribing the singing voice into music notes is challenging due to pitch fluctuations such as portamenti and vibratos. This paper presents a probabilistic transcription method for monophonic sung melodies that explicitly accounts for these local pitch fluctuations. In the hierarchical Hidden Markov Model (HMM), an upper-level ergodic HMM handles the transitions between notes, and a lower-level left-to-right HMM handles the intra- and inter-note pitch fluctuations. The lower-level HMM employs the pitch dynamic model, which explicitly expresses the pitch curve characteristics as the observation likelihood over f0and Δf0using a compact parametric distribution. A histogram-based tuning frequency estimation method, and some post-processing heuristics to separate merged notes and to allocate spuriously detected short notes, improve the note recognition performance. With model parameters that support intuitions about singing behavior, the proposed method obtained encouraging results when evaluated on a published monophonic sung melody dataset, and compared with state-of-the-art methods.

Users of video-sharing sites often search for derivative works of music, such as live versions, covers, and remixes. Audio and video content are both important for retrieval: “karaoke” specifies audio content (instrumental version) and video content (animated lyrics). Although YouTube's text search is fairly reliable, many search results do not match the exact query. We introduce an algorithm to classify YouTube videos by category of derivative work. Based on a standard pipeline for video-based genre classification, it combines search, text, and video features with a novel set of audio features derived from audio fingerprints. A baseline approach is outperformed by the search and text features alone, and combining these with video and audio features performs best of all, reducing the audio content error rate from 25% to 15%.

This study proposes a computational model of the discourse segments in lyrics to understand and to model the structure of lyrics. To test our hypothesis that discourse segmentations in lyrics strongly correlate with repeated patterns, we conduct the first large-scale corpus study on discourse segments in lyrics. Next, we propose the task to automatically identify segment boundaries in lyrics and train a logistic regression model for the task with the repeated pattern and textual features. The results of our empirical experiments illustrate the significance of capturing repeated patterns in predicting the boundaries of discourse segments in lyrics.

Data mining tasks such as music indexing, information retrieval, and similarity search, require an understanding of how listeners process music internally. Many algorithms for automatically analyzing the structure of recorded music assume that a large change in one or another musical feature suggests a section boundary. However, this assumption has not been tested: while our understanding of how listeners segment melodies has advanced greatly in the past decades, little is known about how this process works with more complex, full-textured pieces of music, or how stable this process is across genres. Knowing how these factors affect how boundaries are perceived will help researchers to judge the viability of algorithmic approaches with different corpora of music. We present a statistical analysis of a large corpus of recordings whose formal structure was annotated by expert listeners. We find that the acoustic properties of boundaries in these recordings corroborate findings of previous perceptual experiments. Nearly all boundaries correspond to peaks in novelty functions, which measure the rate of change of a musical feature at a particular time scale. Moreover, most of these boundaries match peaks in novelty for several features at several time scales. We observe that the boundary-novelty relationship can vary with listener, time scale, genre, and musical feature. Finally, we show that a boundary profile derived from a collection of novelty functions correlates with the estimated salience of boundaries indicated by listeners.

To identify repeated patterns and contrasting sections in music, it is common to use self-similarity matrices (SSMs) to visualize and estimate structure. We introduce a novel application for SSMs derived from audio recordings: using them to learn about the potential reasoning behind a listener's annotation. We use SSMs generated by musically-motivated audio features at various timescales to represent contributions to a structural annotation. Since a listener's attention can shift among musical features (e.g., rhythm, timbre, and harmony) throughout a piece, we further break down the SSMs into section-wise components and use quadratic programming (QP) to minimize the distance between a linear sum of these components and the annotated description. We posit that the optimal section-wise weights on the feature components may indicate the features to which a listener attended when annotating a piece, and thus may help us to understand why two listeners disagreed about a piece's structure. We discuss some examples that substantiate the claim that feature relevance varies throughout a piece, using our method to investigate differences between listeners' interpretations, and lastly propose some variations on our method.