Catalogue Search | MBRL
Are you sure you want to remove the book from the shelf?
{{itemTitle}}
422
result(s) for
"Trummer"
Sort by:
The city as a technical being : on the mode of existence of architecture
The city is the largest human artifact. It is made by us, yet simultaneously it makes us, as well as all other nonhuman entities. The particular discourse to which this book on the city contributes is the discipline of architecture. It explores a simple question: How does the city effect the mode of existence of its buildings?
Book
Generating highly customizable python code for data processing with large language models
CARD (Coding Assistant for Relational Data analysis) generates Python code that processes relational queries on raw data. Users can customize generated code via natural language instructions, e.g., by instructing the system to use specific libraries or produce certain output. Internally, CARD uses large language models such as GPT-4o to synthesize code. CARD automatically constructs prompts describing code generation tasks to the language models. Those prompts contain information on data format, customization requirements, as well as processing plans, generated by CARD’s scenario-specific query planner. CARD automatically tests generated code by comparing its output to the output of a reference SQL engine. In case of inconsistencies, CARD re-generates code with a certain degree of randomization. Furthermore, CARD can automatically generate libraries of code samples for specific customization scenarios in a pre-processing step, leveraging those samples at run time for few-shot learning. The experiments show that CARD generates accurate code in the vast majority of scenarios. Furthermore, current trends in language models are likely to benefit CARD’s performance in the future.
Journal Article
Generating highly customizable python code for data processing with large language models
CARD (Coding Assistant for Relational Data analysis) generates Python code that processes relational queries on raw data. Users can customize generated code via natural language instructions, e.g., by instructing the system to use specific libraries or produce certain output. Internally, CARD uses large language models such as GPT-4o to synthesize code. CARD automatically constructs prompts describing code generation tasks to the language models. Those prompts contain information on data format, customization requirements, as well as processing plans, generated by CARD’s scenario-specific query planner. CARD automatically tests generated code by comparing its output to the output of a reference SQL engine. In case of inconsistencies, CARD re-generates code with a certain degree of randomization. Furthermore, CARD can automatically generate libraries of code samples for specific customization scenarios in a pre-processing step, leveraging those samples at run time for few-shot learning. The experiments show that CARD generates accurate code in the vast majority of scenarios. Furthermore, current trends in language models are likely to benefit CARD’s performance in the future.
Journal Article
The European Union needs a policy and strategy to secure access to healthcare for undocumented migrants
Creating an evidence base to support policy and practice should be an urgent objective for the research and policy making communities, argues Ursula Trummer
Journal Article
DB-BERT: making database tuning tools “read” the manual
DB-BERT is a database tuning tool that exploits information gained via natural language analysis of manuals and other relevant text documents. It uses text to identify database system parameters to tune as well as recommended parameter values. DB-BERT applies large, pre-trained language models (specifically, the BERT model) for text analysis. During an initial training phase, it fine-tunes model weights in order to translate natural language hints into recommended settings. At run time, DB-BERT learns to aggregate, adapt, and prioritize hints to achieve optimal performance for a specific database system and benchmark. Both phases are iterative and use reinforcement learning to guide the selection of tuning settings to evaluate (penalizing settings that the database system rejects while rewarding settings that improve performance). In our experiments, we leverage hundreds of text documents about database tuning as input for DB-BERT. We compare DB-BERT against various baselines, considering different benchmarks (TPC-C and TPC-H), metrics (throughput and run time), as well as database systems (PostgreSQL and MySQL). The experiments demonstrate clearly that DB-BERT benefits from combining general information about database tuning, mined from text documents, with scenario-specific insights, gained via trial runs. The full source code of DB-BERT is available online at
https://itrummer.github.io/dbbert/
.
Journal Article
Migrant workers in European informal health care settings
Care for the elderly in Europe increasingly depends on migrant carers, both in institutionalised and informal care settings. For the latter, the vulnerability of migrant carers working in private households has been described in several studies. Austria is a EU member state that made attempts to legalise the grey labour market of home care by several legislative changes, providing a framework of regulations to foster inclusion of migrant care workers into schemes of social protection. It can serve as an example how attempts of inclusion also make room for in-built mechanisms of exclusion, and how public health issues are related to economic considerations of a “care market”, with migrant workers as important, yet rather voiceless, stakeholders.
Journal Article
Carbohydrate microarrays for the recognition of cross-reactive molecular markers of microbes and host cells
We describe here the development of a carbohydrate-based microarray to extend the scope of biomedical research on carbohydrate-mediated molecular recognition and anti-infection responses. We have demonstrated that microbial polysaccharides can be immobilized on a surface-modified glass slide without chemical conjugation. With this procedure, a large repertoire of microbial antigens (∼20,000 spots) can be patterned on a single micro-glass slide, reaching the capacity to include most common pathogens. Glycoconjugates of different structural characteristics are shown here to be applicable for microarray fabrication, extending the repertoires of diversity and complexity of carbohydrate microarrays. The printed microarrays can be air-dried and stably stored at room temperature for long periods of time. In addition, the system is highly sensitive, allowing simultaneous detection of a broad spectrum of antibody specificities with as little as a few microliters of serum specimen. Finally, the potential of carbohydrate microarrays is demonstrated by the discovery of previously undescribed cellular markers, Dex-Ids.
Journal Article
Two distinct modes of DNMT1 recruitment ensure stable maintenance DNA methylation
Stable inheritance of DNA methylation is critical for maintaining differentiated phenotypes in multicellular organisms. We have recently identified dual mono-ubiquitylation of histone H3 (H3Ub2) by UHRF1 as an essential mechanism to recruit DNMT1 to chromatin. Here, we show that PCNA-associated factor 15 (PAF15) undergoes UHRF1-dependent dual mono-ubiquitylation (PAF15Ub2) on chromatin in a DNA replication-coupled manner. This event will, in turn, recruit DNMT1. During early S-phase, UHRF1 preferentially ubiquitylates PAF15, whereas H3Ub2 predominates during late S-phase. H3Ub2 is enhanced under PAF15 compromised conditions, suggesting that H3Ub2 serves as a backup for PAF15Ub2. In mouse ES cells, loss of PAF15Ub2 results in DNA hypomethylation at early replicating domains. Together, our results suggest that there are two distinct mechanisms underlying replication timing-dependent recruitment of DNMT1 through PAF15Ub2 and H3Ub2, both of which are prerequisite for high fidelity DNA methylation inheritance.
Ubiquitylation of histone H3 (H3Ub2) by UHRF1 recruits DNMT1 to chromatin, which is essential for DNA methylation inheritance. Here, the authors provide evidence that there are two distinct mechanisms underlying replication timing-dependent recruitment of DNMT1 through PAF15Ub2 and H3Ub2, both of which are required for high fidelity DNA methylation inheritance.
Journal Article
The Szegö kernel and oblique projections: conformal mapping of non-smooth regions
The method of Kerzman and Trummer (
J. Comp. Appl. Math.
14, 111–123, 1986) for computing the Riemann mapping function of a smooth domain is extended to include the case of simply connected convex regions with corners, in particular convex polygons. The connection between the Szegö kernel and the Riemann mapping function is classical. The integral equation in Kerzman and Trummer (
J. Comp. Appl. Math.
14, 111–123, 1986) that determines the Szegö kernel is no longer defined in the presence of corners. We modify the equation by using new oblique projections. This approach is equivalent to employing preliminary mappings.
Journal Article