When an LLM is apprehensive about its answers -- and when its uncertainty is justified

Uncertainty estimation is crucial for evaluating Large Language Models (LLMs), particularly in high-stakes domains where incorrect answers result in significant consequences. Numerous approaches consider this problem, while focusing on a specific type of uncertainty, ignoring others. We investigate what estimates, specifically token-wise entropy and model-as-judge (MASJ), would work for multiple-choice question-answering tasks for different question topics. Our experiments consider three LLMs: Phi-4, Mistral, and Qwen of different sizes from 1.5B to 72B and $14$ topics. While MASJ performs similarly to a random error predictor, the response entropy predicts model error in knowledge-dependent domains and serves as an effective indicator of question difficulty: for biology ROC AUC is $0.73$. This correlation vanishes for the reasoning-dependent domain: for math questions ROC-AUC is $0.55$. More principally, we found out that the entropy measure required a reasoning amount. Thus, data-uncertainty related entropy should be integrated within uncertainty estimates frameworks, while MASJ requires refinement. Moreover, existing MMLU-Pro samples are biased, and should balance required amount of reasoning for different subdomains to provide a more fair assessment of LLMs performance.

Concealed Adversarial attacks on neural networks for sequential data

The emergence of deep learning led to the broad usage of neural networks in the time series domain for various applications, including finance and medicine. While powerful, these models are prone to adversarial attacks: a benign targeted perturbation of input data leads to significant changes in a classifier's output. However, formally small attacks in the time series domain become easily detected by the human eye or a simple detector model. We develop a concealed adversarial attack for different time-series models: it provides more realistic perturbations, being hard to detect by a human or model discriminator. To achieve this goal, the proposed adversarial attack maximizes an aggregation of a classifier and a trained discriminator loss. To make the attack stronger, we also propose a training procedure for a discriminator that provides broader coverage of possible attacks. Extensive benchmarking on six UCR time series datasets across four diverse architectures - including recurrent, convolutional, state-space, and transformer-based models - demonstrates the superiority of our attack for a concealability-efficiency trade-off. Our findings highlight the growing challenge of designing robust time series models, emphasizing the need for improved defenses against realistic and effective attacks.

Normalizing self-supervised learning for provably reliable Change Point Detection

Change point detection (CPD) methods aim to identify abrupt shifts in the distribution of input data streams. Accurate estimators for this task are crucial across various real-world scenarios. Yet, traditional unsupervised CPD techniques face significant limitations, often relying on strong assumptions or suffering from low expressive power due to inherent model simplicity. In contrast, representation learning methods overcome these drawbacks by offering flexibility and the ability to capture the full complexity of the data without imposing restrictive assumptions. However, these approaches are still emerging in the CPD field and lack robust theoretical foundations to ensure their reliability. Our work addresses this gap by integrating the expressive power of representation learning with the groundedness of traditional CPD techniques. We adopt spectral normalization (SN) for deep representation learning in CPD tasks and prove that the embeddings after SN are highly informative for CPD. Our method significantly outperforms current state-of-the-art methods during the comprehensive evaluation via three standard CPD datasets.

Collusion Detection with Graph Neural Networks

Collusion is a complex phenomenon in which companies secretly collaborate to engage in fraudulent practices. This paper presents an innovative methodology for detecting and predicting collusion patterns in different national markets using neural networks (NNs) and graph neural networks (GNNs). GNNs are particularly well suited to this task because they can exploit the inherent network structures present in collusion and many other economic problems. Our approach consists of two phases: In Phase I, we develop and train models on individual market datasets from Japan, the United States, two regions in Switzerland, Italy, and Brazil, focusing on predicting collusion in single markets. In Phase II, we extend the models' applicability through zero-shot learning, employing a transfer learning approach that can detect collusion in markets in which training data is unavailable. This phase also incorporates out-of-distribution (OOD) generalization to evaluate the models' performance on unseen datasets from other countries and regions. In our empirical study, we show that GNNs outperform NNs in detecting complex collusive patterns. This research contributes to the ongoing discourse on preventing collusion and optimizing detection methodologies, providing valuable guidance on the use of NNs and GNNs in economic applications to enhance market fairness and economic welfare.

Gallery-Aware Uncertainty Estimation For Open-Set Face Recognition

Accurately estimating image quality and model robustness improvement are critical challenges in unconstrained face recognition, which can be addressed through uncertainty estimation via probabilistic face embeddings. Previous research mainly focused on uncertainty estimation in face verification, leaving the open-set face recognition task underexplored. In open-set face recognition, one seeks to classify an image, which could also be unknown. Here, the low variance of probabilistic embedding does not imply a low error probability: an image embedding could be close to several classes in a gallery, thus yielding high uncertainty. We propose a method aware of two sources of ambiguity in the open-set recognition system: (1) the gallery uncertainty caused by overlapping classes and (2) the uncertainty of the face embeddings. To detect both types, we use a Bayesian probabilistic model of embedding distribution, which provides a principled uncertainty estimate. Challenging open-set face recognition datasets, such as IJB-C, serve as a testbed for our method. We also propose a new open-set recognition protocol for whale and dolphin identification. The proposed approach better identifies recognition errors than uncertainty estimation methods based solely on image quality.

Uniting contrastive and generative learning for event sequences models

High-quality representation of transactional sequences is vital for modern banking applications, including risk management, churn prediction, and personalized customer offers. Different tasks require distinct representation properties: local tasks benefit from capturing the client's current state, while global tasks rely on general behavioral patterns. Previous research has demonstrated that various self-supervised approaches yield representations that better capture either global or local qualities. This study investigates the integration of two self-supervised learning techniques - instance-wise contrastive learning and a generative approach based on restoring masked events in latent space. The combined approach creates representations that balance local and global transactional data characteristics. Experiments conducted on several public datasets, focusing on sequence classification and next-event type prediction, show that the integrated method achieves superior performance compared to individual approaches and demonstrates synergistic effects. These findings suggest that the proposed approach offers a robust framework for advancing event sequences representation learning in the financial sector.

COTODE: COntinuous Trajectory neural Ordinary Differential Equations for modelling event sequences

Observation of the underlying actors that generate event sequences reveals that they often evolve continuously. Most modern methods, however, tend to model such processes through at most piecewise-continuous trajectories. To address this, we adopt a way of viewing events not as standalone phenomena but instead as observations of a Gaussian Process, which in turn governs the actor's dynamics. We propose integrating these obtained dynamics, resulting in a continuous-trajectory modification of the widely successful Neural ODE model. Through Gaussian Process theory, we were able to evaluate the uncertainty in an actor's representation, which arises from not observing them between events. This estimate led us to develop a novel, theoretically backed negative feedback mechanism. Empirical studies indicate that our model with Gaussian process interpolation and negative feedback achieves state-of-the-art performance, with improvements up to 20% AUROC against similar architectures.

Universal representations for financial transactional data: embracing local, global, and external contexts

Effective processing of financial transactions is essential for banking data analysis. However, in this domain, most methods focus on specialized solutions to stand-alone problems instead of constructing universal representations suitable for many problems. We present a representation learning framework that addresses diverse business challenges. We also suggest novel generative models that account for data specifics, and a way to integrate external information into a client's representation, leveraging insights from other customers' actions. Finally, we offer a benchmark, describing representation quality globally, concerning the entire transaction history; locally, reflecting the client's current state; and dynamically, capturing representation evolution over time. Our generative approach demonstrates superior performance in local tasks, with an increase in ROC-AUC of up to 14\% for the next MCC prediction task and up to 46\% for downstream tasks from existing contrastive baselines. Incorporating external information improves the scores by an additional 20\%.

Diversity-Aware Ensembling of Language Models Based on Topological Data Analysis

Ensembles are important tools for improving the performance of machine learning models. In cases related to natural language processing, ensembles boost the performance of a method due to multiple large models available in open source. However, existing approaches mostly rely on simple averaging of predictions by ensembles with equal weights for each model, ignoring differences in the quality and conformity of models. We propose to estimate weights for ensembles of NLP models using not only knowledge of their individual performance but also their similarity to each other. By adopting distance measures based on Topological Data Analysis (TDA), we improve our ensemble. The quality improves for both text classification accuracy and relevant uncertainty estimation.

From Variability to Stability: Advancing RecSys Benchmarking Practices

In the rapidly evolving domain of Recommender Systems (RecSys), new algorithms frequently claim state-of-the-art performance based on evaluations over a limited set of arbitrarily selected datasets. However, this approach may fail to holistically reflect their effectiveness due to the significant impact of dataset characteristics on algorithm performance. Addressing this deficiency, this paper introduces a novel benchmarking methodology to facilitate a fair and robust comparison of RecSys algorithms, thereby advancing evaluation practices. By utilizing a diverse set of $30$ open datasets, including two introduced in this work, and evaluating $11$ collaborative filtering algorithms across $9$ metrics, we critically examine the influence of dataset characteristics on algorithm performance. We further investigate the feasibility of aggregating outcomes from multiple datasets into a unified ranking. Through rigorous experimental analysis, we validate the reliability of our methodology under the variability of datasets, offering a benchmarking strategy that balances quality and computational demands. This methodology enables a fair yet effective means of evaluating RecSys algorithms, providing valuable guidance for future research endeavors.