Θεόδωρος Γιαννακόπουλος

Panagiotis Koromilas, Theodoros Giannakopoulos, Mihalis A. Nicolaou et al.

Supervised classification has a theoretical optimum, Neural Collapse (NC), yet neither of its two dominant paradigms reaches it in practice. Cross entropy (CE) leaves radial degrees of freedom unconstrained and converges to a degenerate geometry, while supervised contrastive learning (SCL) drives features toward NC during pretraining but discards this structure in a post hoc linear probing phase. We show that both paradigms are different appearances of the same method, prototype contrast on the unit hypersphere, and that closing the gap requires fixing each at its specific point of failure. From the CE side, we propose NTCE and NONL, two normalized losses that import contrastive optimization's missing ingredients into classifier learning: a large effective negative set and decoupled alignment and uniformity terms. From the SCL side, we prove that SCL's objective already optimizes throughout training for a principled classifier whose weights are the class mean embeddings, making linear probing both redundant and harmful. Empirically, on four benchmarks including ImageNet-1K, NTCE and NONL surpass CE accuracy, closely approximate NC ($\geq 95\%$), and match CE's converged NC on 4/5 metrics in under $7.5\%$ of its iterations, while SCL with fixed prototypes matches linear probing without the hours-long classifier training phase. The learned geometry yields $+5.5\%$ mean relative improvement in transfer learning, up to $+8.7\%$ under severe class imbalance, and lower mCE on ImageNet-C, recasting supervised learning as prototype learning on the hypersphere, with NC reached by design on both paths.

Evangelia Gkritzali, Panagiotis Kaliosis, Sofia Galanaki et al.

In the vast majority of the academic and scientific domains, LaTeX has established itself as the de facto standard for typesetting complex mathematical equations and formulae. However, LaTeX's complex syntax and code-like appearance present accessibility barriers for individuals with disabilities, as well as those unfamiliar with coding conventions. In this paper, we present a novel solution to this challenge through the development of a novel speech-to-LaTeX equations system specifically designed for the Greek language. We propose an end-to-end system that harnesses the power of Automatic Speech Recognition (ASR) and Natural Language Processing (NLP) techniques to enable users to verbally dictate mathematical expressions and equations in natural language, which are subsequently converted into LaTeX format. We present the architecture and design principles of our system, highlighting key components such as the ASR engine, the LLM-based prompt-driven equations generation mechanism, as well as the application of a custom evaluation metric employed throughout the development process. We have made our system open source and available at https://github.com/magcil/greek-speech-to-math.

Panagiotis Koromilas, Efthymios Georgiou, Giorgos Bouritsas et al.

Contrastive Learning (CL), a leading paradigm in Self-Supervised Learning (SSL), typically relies on pairs of data views generated through augmentation. While multiple augmentations per instance (more than two) improve generalization in supervised learning, current CL methods handle additional views suboptimally by simply aggregating different pairwise objectives. This approach suffers from four critical limitations: (L1) it utilizes multiple optimization terms per data point resulting to conflicting objectives, (L2) it fails to model all interactions across views and data points, (L3) it inherits fundamental limitations (e.g. alignment-uniformity coupling) from pairwise CL losses, and (L4) it prevents fully realizing the benefits of increased view multiplicity observed in supervised settings. We address these limitations through two novel loss functions: MV-InfoNCE, which extends InfoNCE to incorporate all possible view interactions simultaneously in one term per data point, and MV-DHEL, which decouples alignment from uniformity across views while scaling interaction complexity with view multiplicity. Both approaches are theoretically grounded - we prove they asymptotically optimize for alignment of all views and uniformity, providing principled extensions to multi-view contrastive learning. Our empirical results on ImageNet1K and three other datasets demonstrate that our methods consistently outperform existing multi-view approaches and effectively scale with increasing view multiplicity. We also apply our objectives to multimodal data and show that, in contrast to other contrastive objectives, they can scale beyond just two modalities. Most significantly, ablation studies reveal that MV-DHEL with five or more views effectively mitigates dimensionality collapse by fully utilizing the embedding space, thereby delivering multi-view benefits observed in supervised learning.

Christos Nikou, Theodoros Giannakopoulos

Recent advances in song identification leverage deep neural networks to learn compact audio fingerprints directly from raw waveforms. While these methods perform well under controlled conditions, their accuracy drops significantly in real-world scenarios where the audio is captured via mobile devices in noisy environments. In this paper, we introduce a novel evaluation protocol designed to better reflect such real-world conditions. We generate three recordings of the same audio, each with increasing levels of noise, captured using a mobile device's microphone. Our results reveal a substantial performance drop for two state-of-the-art CNN-based models under this protocol, compared to previously reported benchmarks. Additionally, we highlight the critical role of the augmentation pipeline during training with contrastive loss. By introduction low pass and high pass filters in the augmentation pipeline we significantly increase the performance of both systems in our proposed evaluation. Furthermore, we develop a transformer-based model with a tailored projection module and demonstrate that transferring knowledge from a semantically relevant domain yields a more robust solution. The transformer architecture outperforms CNN-based models across all noise levels, and query durations. In low noise conditions it achieves 47.99% for 1-sec queries, and 97% for 10-sec queries in finding the correct song, surpassing by 14%, and by 18.5% the second-best performing model, respectively, Under heavy noise levels, we achieve a detection rate 56.5% for 15-second query duration. All experiments are conducted on public large-scale dataset of over 100K songs, with queries matched against a database of 56 million vectors.

Georgios Paraskevopoulos, Efthymios Tzinis, Nikolaos Ellinas et al.

We examine the use of linear and non-linear dimensionality reduction algorithms for extracting low-rank feature representations for speech emotion recognition. Two feature sets are used, one based on low-level descriptors and their aggregations (IS10) and one modeling recurrence dynamics of speech (RQA), as well as their fusion. We report speech emotion recognition (SER) results for learned representations on two databases using different classification methods. Classification with low-dimensional representations yields performance improvement in a variety of settings. This indicates that dimensionality reduction is an effective way to combat the curse of dimensionality for SER. Visualization of features in two dimensions provides insight into discriminatory abilities of reduced feature sets.

Konstantinos Bougiatiotis, Theodore Giannakopoulos

In this paper we examine the ability of low-level multimodal features to extract movie similarity, in the context of a content-based movie recommendation approach. In particular, we demonstrate the extraction of multimodal representation models of movies, based on textual information from subtitles, as well as cues from the audio and visual channels. With regards to the textual domain, we emphasize our research in topic modeling of movies based on their subtitles, in order to extract topics that discriminate between movies. Regarding the visual domain, we focus on the extraction of semantically useful features that model camera movements, colors and faces, while for the audio domain we adopt simple classification aggregates based on pretrained models. The three domains are combined with static metadata (e.g. directors, actors) to prove that the content-based movie similarity procedure can be enhanced with low-level multimodal information. In order to demonstrate the proposed content representation approach, we have built a small dataset of 160 widely known movies. We assert movie similarities, as propagated by the individual modalities and fusion models, in the form of recommendation rankings. Extensive experimentation proves that all three low-level modalities (text, audio and visual) boost the performance of a content-based recommendation system, compared to the typical metadata-based content representation, by more than $50\%$ relative increase. To our knowledge, this is the first approach that utilizes a wide range of features from all involved modalities, in order to enhance the performance of the content similarity estimation, compared to the metadata-based approaches.

Sergios Petridis, Theodoros Giannakopoulos, Costantine D. Spyropoulos

Unobtrusive every day health monitoring can be of important use for the elderly population. In particular, pupil size may be a valuable source of information, since, apart from pathological cases, it can reveal the emotional state, the fatigue and the ageing. To allow for unobtrusive monitoring to gain acceptance, one should seek for efficient methods of monitoring using com- mon low-cost hardware. This paper describes a method for monitoring pupil sizes using a common web camera in real time. Our method works by first detecting the face and the eyes area. Subsequently, optimal iris and sclera location and radius, modelled as ellipses, are found using efficient filtering. Finally, the pupil center and radius is estimated by optimal filtering within the area of the iris. Experimental result show both the efficiency and the effectiveness of our approach.