Our Work

Ilmarinen selects a neural architecture -- its computational contract (arena), its primitive operations, and its width and depth -- for a dataset by minimizing a single description-length objective, w…

Ilmarinen selects a neural architecture -- its computational contract (arena), its primitive operations, and its width and depth -- for a dataset by minimizing a single description-length objective, with every complexity charge derived as a code length rather than a hand-set constant.

Let $\Phi:M_d(\C)\to M_d(\C)$ be a completely positive map with a transpose-symmetric Kraus representation. Its Choi operator determines the rank-two constant \[ \beta_2(\Phi) =\frac12\sup_{\substack{…

Let $\Phi:M_d(\C)\to M_d(\C)$ be a completely positive map with a transpose-symmetric Kraus representation. Its Choi operator determines the rank-two constant \[ \beta_2(\Phi) =\frac12\sup_{\substack{\norm z=1\\ \SR(z)\le2}} \ip z{J(\Phi)z}. \] We prove that, for every $r\ge1$, \[ \Phi\circ\tau+(r-1)\beta_2(\Phi) \left(\Delta-\frac1r\id\right) \] is $r$-positive, where $\tau$ is transposition and $\Delta(X)=\Tr(X)I$. The proof turns the negative spectrum of a partially transposed rank-one operator into the edges of the complete graph $K_r$. Regrouping by vertices costs its degree $r-1$, rather than its $\binom r2$ edges. Transpose symmetry protects one direction and is exactly what produces the term $-\id/r$. For $\Phi=\Lambda_U\otimes\Lambda_V$, with $\Lambda_E(X)=\Tr(X)I-X^T$, a sharp double-skew singular-value estimate gives $\beta_2\le1$, with equality when both local dimensions are at least two. A map identity then yields, for every operator $C$ of rank at most $r$ on $U\otimes V$, \[ \norm{\Tr_U C}_2^2+\norm{\Tr_V C}_2^2 \le r\norm C_2^2+\frac1r\abs{\Tr C}^2, \] with dimension-uniform coefficients that are sharp whenever one local dimension is at least $r$ and the other is at least two. A second family, built from the edges of an arbitrary graph, has the same local constant and attains the universal coefficient $r-1$ whenever the graph contains an $r$-clique. The Choi lifting chain through the partial-trace and graph results, together with its sharpness witnesses, is formalized in Lean~4.

The relevance of the issues developed by the author is largely due to the following circumstances. Firstly, by improving the modern control and supervisory activities of public authorities, sectoral m…

The relevance of the issues developed by the author is largely due to the following circumstances. Firstly, by improving the modern control and supervisory activities of public authorities, sectoral management, and public institutions. In this regard, taking into account the historical experience of the implementation of the control function has a certain value in terms of identifying the root principles of the control activities of various subjects, in reflecting the continuity of common and different views on the importance of control in the past and in modern reality. Secondly, the importance of increasing scientific knowledge in domestic historical and legal science, in particular in the study of the functioning of domestic law enforcement agencies. The main purpose of the work is to identify the essence of control both in the doctrinal aspect in general and in the historical aspect, in particular, as well as in relation to the domestic penitentiary system of the stated historical period, as well as to establish a periodization of the evolution of the organizational and legal foundations of control in this area. The article examines the problems of various approaches to understanding the essence of control, as well as to the periodization of the formation and development of the organizational and legal foundations of control in the Russian penitentiary system of the pre-revolutionary period. The methods of scientific knowledge used are dialectics, analysis, synthesis, method of legal hermeneutics, historical and typological method, comparative legal method. As a result of the research, the author comes to conclusions about the multifaceted nature of the concept of control in the theoretical and legal aspect.: as functions of the state, as forms of legal activity, as guarantees of legality. An analysis of a wide range of historical sources allowed the author to establish the fact that the concept of control was not fixed in the norms of pre-revolutionary imperial legislation, as well as penal legislation, in particular, and to formulate the author’s definition of it. Also, based on the analysis of various approaches to the periodization of the formation and development of public authorities and sectoral management, the author provides the author’s periodization of the organizational and legal foundations of control in the field of the execution of criminal penalties of the specified historical period.

RPT15 Demonstration Subset: This record contains a five-subject demonstration subset of RPT15. It provides 300 seconds of single-lead ECG data per subject with manual R-, P-, and T-peak annotations. T…

RPT15 Demonstration Subset: This record contains a five-subject demonstration subset of RPT15. It provides 300 seconds of single-lead ECG data per subject with manual R-, P-, and T-peak annotations. The subset accompanies the manuscript “CardioAnnotate: An Interactive Toolkit for ECG Signal Processing, Annotation, Gaussian Process Correction, and Interoperable Data Exchange” and provides runnable example data under a permanent DOI. It is not the complete RPT15 dataset release. The complete 42-subject archive is currently under review at PhysioNet. A link to the full release will be added to this record after publication. Data contents Each subject is provided at its native sampling rate and at fixed sampling rates of 250 Hz and 512 Hz. Every record contains one canonical ECG lead. Data are distributed in: NPZ format, including the ECG signal and metadata. WFDB format, including a time-ordered annotation stream using N for R peaks, p for P peaks, and t for T peaks. CSV and JSON provenance files identifying the source record, database, lead, native sampling rate, and extracted window. A SHA256SUMS file containing checksums for the distributed files. Record names follow the convention {subject_id}_{release}. Subject identifiers are preserved from the complete RPT15 archive and are therefore not contiguous. For example, 0001_variable in this subset and 0001_variable in the complete archive identify the same subject. See README.md for the complete directory structure, metadata description, subject table, reconstruction instructions, and annotation-adjustment parameters. License and attribution The ECG signals are derived from PhysioNet databases made available under the Open Data Commons Attribution License v1.0. This demonstration subset is redistributed under the same license. See LICENSE.txt for the applicable license terms and ATTRIBUTION.md for the source databases, source links, attribution notices, and required PhysioNet citation. Citation When using this demonstration subset, please cite: Mojtahed H. RPT15 Sample: a 5-subject demonstration subset of single-lead ECG with R, P, and T peak annotations at various sampling rates [Dataset]. Zenodo, 2026. https://doi.org/10.5281/zenodo.21652810 Mojtahed H. CardioAnnotate: Browser-Based ECG Annotation, RR Correction, HRV Analysis, and Export [Computer software]. Zenodo, 2026. https://doi.org/10.5281/zenodo.20820831 Please also cite the source databases identified in ATTRIBUTION.md. This subset was produced with CardioAnnotate. The publication reference for the accompanying manuscript will be added when it becomes available. Related resources CardioAnnotate toolkit live website: Launch CardioAnnotate Project website: Visit the project website

This aggregated dataset consists of a corpus of qualitative oral history interviews collected within the Horizon Europe project DanceMap (Grant No. 101177556). With the aim to document diverse approac…

This aggregated dataset consists of a corpus of qualitative oral history interviews collected within the Horizon Europe project DanceMap (Grant No. 101177556). With the aim to document diverse approaches to and understandings of contemporary dance heritage, the interviews are part of the fieldwork conducted directly with representatives of the project’s partner institutions in Cologne, Sofia, Skopje, Belgrade, Lisbon, Ljubljana, Brussels, and Leuven. By bringing together the voices of curators, academic researchers, artists, cultural producers and others, the collection offers a diversified insight into the institutional realities of as well as the critical discourse surrounding dance heritage practices in different national and political contexts across Europe. The dataset is intentionally designed to offer deep, contextual narratives suitable for qualitative research. It is conceived as a primary or “raw” corpus of full text transcripts rather than a closed collection of terms or pre-determined concepts, maximising its potential for secondary data reuse both within and beyond the DanceMap framework. Geographically and institutionally, the dataset represents a unified effort across 9 partner organisations spanning 8 European nations, resulting in a collection of 16 interviews with 25 interviewees based on 27 hours and 32 minutes of recorded material. Access and documentation are provided via three core root-level files: a synthesising README.txt file acting as the technical index and directory guide, alongside a master data matrix (participants_metadata.csv) for automated cross-referencing. 

This study explores enhanced oil recovery (EOR) strategies, with a focus on carbonate reservoirs constituting over 60% of global oil discoveries. While “smart water” injection proves effec…

This study explores enhanced oil recovery (EOR) strategies, with a focus on carbonate reservoirs constituting over 60% of global oil discoveries. While “smart water” injection proves effective in EOR for carbonate reservoirs, offshore application challenges arise due to impractical volumes for injection. To address this, we propose a novel continuous injection approach, systematically investigating it on a laboratory scale using the Iranian offshore reservoir, Sivand. Thirty-six contact angle tests and twelve flooding experiments are meticulously conducted, with key ions, potassium, and sulfate, playing pivotal roles. Optimal wettability alteration is observed at 4 times potassium ion concentration in 0–2 times sulfate concentrations, driven by ionic strength and charge interactions. Conversely, at 3–5 times sulfate concentrations, the optimal contact angle shifts to 2 times potassium ion concentration, suggesting a mechanism change linked to increasing sulfate ion ionicity. A significant wettability alteration, evidenced by a 132.8° decrease, occurs in seawater with a twofold concentration of potassium ions and a fivefold concentration of sulfate ions. Micromodel experiments introduce an innovative alternation of smart water and seawater injections. The first scenario, smart water followed by seawater injection, reveals negligible post-seawater injection oil recovery changes. In contrast, the second scenario yields a maximum recovery of 7.9%. The first scenario, however, boasts superior overall sweep efficacy, reaching approximately 43%. This research expands understanding of smart water and seawater injection in EOR, presenting a viable solution for optimizing offshore carbonate reservoir recovery. The insights contribute to evolving EOR methodologies, emphasizing tailored strategies for varying reservoir conditions.

The relevance of the study is determined by the fact that the current version of Article 203 of the Criminal Code of the Russian Federation retains a private detective as a special subject of excess o…

The relevance of the study is determined by the fact that the current version of Article 203 of the Criminal Code of the Russian Federation retains a private detective as a special subject of excess of authority, but does not sufficiently reflect the informational nature of modern abuses in private investigation. The main purpose of the article is to develop a legislative model of a new qualified element of Article 203 intended to assess unlawful collection, receipt, storage, use, transfer or dissemination of legally protected information by a private detective in the course of private detective activity. The article examines the lenient sanction of Part 1 of Article 203, the predominantly violent and security-oriented focus of the current Part 2, the inadmissibility of replacing the professional status of a private detective with the feature of official position, and the relationship between the proposed rule and special information crimes. Particular attention is paid to the limits of cumulative qualification and to the need for explanations of the Plenum of the Supreme Court concerning informational consequences of excess of authority. The study uses formal legal, systemic structural, law-enforcement methods and the method of criminal law modeling. The results are expressed in a draft qualified element, its sanction, crime category and qualification algorithm. The article concludes that a private detective should remain within Article 203, provided that the law specifies the independent professional harm caused to the licensed private investigation regime

Material shortages, late deliveries and mis-timed procurement remain among the most persistent causes of schedule overrun in building and infrastructure projects. Traditional material planning still d…

Material shortages, late deliveries and mis-timed procurement remain among the most persistent causes of schedule overrun in building and infrastructure projects. Traditional material planning still depends on static lead times taken from vendor quotations, bills of quantity that are frozen early, and uniform inventory buffers that treat a truckload of cement and an imported glazing package as if they carried the same risk. This chapter examines how artificial intelligence (AI), machine learning and explainable AI (XAI) can be combined into a single predictive material planning framework, and what that combination actually buys a project team in terms of avoided delay. The argument is developed through a synthesis of the delay-causation and construction supply chain literature, three strategic analysis lenses (PESTEL, Porter's Five Forces and SWOT), and an empirical illustration built on a 1,850-record material procurement dataset covering twelve material categories. Gradient-boosted and random forest models predicted realised delivery lead time with a mean absolute error of 3.61 days, against 6.02 days for the quotation-based baseline, a reduction of roughly forty per cent. Schedule-critical orders were identified with an ROC-AUC of 0.885. SHAP analysis showed that supplier on-time history, planned inventory buffer, transport distance and internal approval cycle time dominated the predictions, while project stage contributed about two per cent of total attribution. A buffer-reallocation simulation demonstrated that redistributing the same total inventory-days according to predicted risk cut unprotected delay exposure by 37.1 per cent without any additional working capital. The chapter contributes an integrated framework linking prediction, explanation and procurement decision-making, and sets out practical guidance for contractors at different levels of digital maturity.