Andrey Kravchenko

Petr Anokhin, Nikita Semenov, Artyom Sorokin et al.

Advancements in the capabilities of Large Language Models (LLMs) have created a promising foundation for developing autonomous agents. With the right tools, these agents could learn to solve tasks in new environments by accumulating and updating their knowledge. Current LLM-based agents process past experiences using a full history of observations, summarization, retrieval augmentation. However, these unstructured memory representations do not facilitate the reasoning and planning essential for complex decision-making. In our study, we introduce AriGraph, a novel method wherein the agent constructs and updates a memory graph that integrates semantic and episodic memories while exploring the environment. We demonstrate that our Ariadne LLM agent, consisting of the proposed memory architecture augmented with planning and decision-making, effectively handles complex tasks within interactive text game environments difficult even for human players. Results show that our approach markedly outperforms other established memory methods and strong RL baselines in a range of problems of varying complexity. Additionally, AriGraph demonstrates competitive performance compared to dedicated knowledge graph-based methods in static multi-hop question-answering.

Mark Rofin, Vladislav Mikhailov, Mikhail Florinskiy et al.

The development of state-of-the-art systems in different applied areas of machine learning (ML) is driven by benchmarks, which have shaped the paradigm of evaluating generalisation capabilities from multiple perspectives. Although the paradigm is shifting towards more fine-grained evaluation across diverse tasks, the delicate question of how to aggregate the performances has received particular interest in the community. In general, benchmarks follow the unspoken utilitarian principles, where the systems are ranked based on their mean average score over task-specific metrics. Such aggregation procedure has been viewed as a sub-optimal evaluation protocol, which may have created the illusion of progress. This paper proposes Vote'n'Rank, a framework for ranking systems in multi-task benchmarks under the principles of the social choice theory. We demonstrate that our approach can be efficiently utilised to draw new insights on benchmarking in several ML sub-fields and identify the best-performing systems in research and development case studies. The Vote'n'Rank's procedures are more robust than the mean average while being able to handle missing performance scores and determine conditions under which the system becomes the winner.

Milena Gazdieva, Petr Mokrov, Litu Rout et al.

Real-world image super-resolution (SR) tasks often do not have paired datasets, which limits the application of supervised techniques. As a result, the tasks are usually approached by unpaired techniques based on Generative Adversarial Networks (GANs), which yield complex training losses with several regularization terms, e.g., content or identity losses. While GANs usually provide good practical performance, they are used heuristically, i.e., theoretical understanding of their behaviour is yet rather limited. We theoretically investigate optimization problems which arise in such models and find two surprising observations. First, the learned SR map is always an optimal transport (OT) map. Second, we theoretically prove and empirically show that the learned map is biased, i.e., it does not actually transform the distribution of low-resolution images to high-resolution ones. Inspired by these findings, we investigate recent advances in neural OT field to resolve the bias issue. We establish an intriguing connection between regularized GANs and neural OT approaches. We show that unlike the existing GAN-based alternatives, these algorithms aim to learn an unbiased OT map. We empirically demonstrate our findings via a series of synthetic and real-world unpaired SR experiments. Our source code is publicly available at https://github.com/milenagazdieva/OT-Super-Resolution.

Ivan Fursov, Alexey Zaytsev, Pavel Burnyshev et al.

Robustness of huge Transformer-based models for natural language processing is an important issue due to their capabilities and wide adoption. One way to understand and improve robustness of these models is an exploration of an adversarial attack scenario: check if a small perturbation of an input can fool a model. Due to the discrete nature of textual data, gradient-based adversarial methods, widely used in computer vision, are not applicable per~se. The standard strategy to overcome this issue is to develop token-level transformations, which do not take the whole sentence into account. In this paper, we propose a new black-box sentence-level attack. Our method fine-tunes a pre-trained language model to generate adversarial examples. A proposed differentiable loss function depends on a substitute classifier score and an approximate edit distance computed via a deep learning model. We show that the proposed attack outperforms competitors on a diverse set of NLP problems for both computed metrics and human evaluation. Moreover, due to the usage of the fine-tuned language model, the generated adversarial examples are hard to detect, thus current models are not robust. Hence, it is difficult to defend from the proposed attack, which is not the case for other attacks.

Ivan Fursov, Alexey Zaytsev, Nikita Kluchnikov et al.

Deep learning models suffer from a phenomenon called adversarial attacks: we can apply minor changes to the model input to fool a classifier for a particular example. The literature mostly considers adversarial attacks on models with images and other structured inputs. However, the adversarial attacks for categorical sequences can also be harmful. Successful attacks for inputs in the form of categorical sequences should address the following challenges: (1) non-differentiability of the target function, (2) constraints on transformations of initial sequences, and (3) diversity of possible problems. We handle these challenges using two black-box adversarial attacks. The first approach adopts a Monte-Carlo method and allows usage in any scenario, the second approach uses a continuous relaxation of models and target metrics, and thus allows usage of state-of-the-art methods for adversarial attacks with little additional effort. Results for money transactions, medical fraud, and NLP datasets suggest that proposed methods generate reasonable adversarial sequences that are close to original ones but fool machine learning models.

Vage Egiazarian, Savva Ignatyev, Alexey Artemov et al.

Constructing high-quality generative models for 3D shapes is a fundamental task in computer vision with diverse applications in geometry processing, engineering, and design. Despite the recent progress in deep generative modelling, synthesis of finely detailed 3D surfaces, such as high-resolution point clouds, from scratch has not been achieved with existing approaches. In this work, we propose to employ the latent-space Laplacian pyramid representation within a hierarchical generative model for 3D point clouds. We combine the recently proposed latent-space GAN and Laplacian GAN architectures to form a multi-scale model capable of generating 3D point clouds at increasing levels of detail. Our evaluation demonstrates that our model outperforms the existing generative models for 3D point clouds.

Luigi Bellomarini, Ruslan R. Fayzrakhmanov, Georg Gottlob et al.

Following the recent successful examples of large technology companies, many modern enterprises seek to build knowledge graphs to provide a unified view of corporate knowledge and to draw deep insights using machine learning and logical reasoning. There is currently a perceived disconnect between the traditional approaches for data science, typically based on machine learning and statistical modelling, and systems for reasoning with domain knowledge. In this paper we present a state-of-the-art Knowledge Graph Management System, Vadalog, which delivers highly expressive and efficient logical reasoning and provides seamless integration with modern data science toolkits, such as the Jupyter platform. We demonstrate how to use Vadalog to perform traditional data wrangling tasks, as well as complex logical and probabilistic reasoning. We argue that this is a significant step forward towards combining machine learning and reasoning in data science.