Boshko Koloski
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Boshko Koloski has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Blaž Škrlj
Jožef Stefan Institute, Ljubljana, Slovenia
11 shared publications - Senja Pollak
Jožef Stefan Institute, Ljubljana, Slovenia
10 shared publications - Nada Lavrač
Jozef Stefan Institute
7 shared publications Senja Pollak
Jožef Stefan Institute and International Postgraduate School, Ljubljana, Slovenia
5 shared publications- Andrej Kastrin
Institute for Biostatistics and Medical Informatics, Faculty of Medicine, University of Ljubljana, 1000 Ljubljana, Slovenia
4 shared publications Nada Lavrač
Jožef Stefan Institute and International Postgraduate School, Ljubljana, Slovenia
4 shared publications
Latest Publications
- Open Access
- 18 September 2026
- Published by Springer Nature in Machine Learning
- Vol. 115 (10) , 225
Incremental Graph Construction Enables Robust Spectral Clustering of Texts
Neighborhood graphs over document embeddings are a common component of text-mining pipelines, supporting topic discovery, deduplication, semi-supervised label propagation, and retrieval-index construction, and they are a critical but fragile step in spectral clustering of text embeddings. On realistic text datasets, standard k -NN graphs can contain many disconnected components at practical sparsity levels (small k ), making spectral clustering degenerate and sensitive to hyperparameters. We introduce an incremental k -NN graph construction algorithm in which each new node is linked to its k nearest previously inserted nodes; this guarantees a connected graph for any k . We provide an inductive proof of connectedness and discuss implications for incremental updates when new documents arrive. We validate the approach on spectral clustering of SentenceTransformer embeddings using Laplacian eigenmaps across eleven sentence- and paragraph-level clustering tasks drawn from six dataset sources in the Massive Text Embedding Benchmark. Compared to standard k -NN graphs, our method outperforms in the low- k regime where disconnected components are prevalent, and matches standard k -NN at larger k . The advantage at low k persists when the standard k -NN graph is repaired with a minimum spanning tree, while our construction avoids the dense distance matrix an exact repair requires and is consequently far cheaper to build, in both time and memory. The code is available on https://github.com/bkolosk1/incremental_clustering_graphs.
- 18 September 2026
- Published by Springer Nature in Machine Learning
- Vol. 115 (10) , 225
Open AccessIncremental Graph Construction Enables Robust Spectral Clustering of Texts
Neighborhood graphs over document embeddings are a common component of text-mining pipelines, supporting topic discovery, deduplication, semi-supervised label propagation, and retrieval-index construction, and they are a critical but fragile step in spectral clustering of text embeddings. On realistic text datasets, standard k -NN graphs can contain many disconnected components at practical sparsity levels (small k ), making spectral clustering degenerate and sensitive to hyperparameters. We introduce an incremental k -NN graph construction algorithm in which each new node is linked to its k nearest previously inserted nodes; this guarantees a connected graph for any k . We provide an inductive proof of connectedness and discuss implications for incremental updates when new documents arrive. We validate the approach on spectral clustering of SentenceTransformer embeddings using Laplacian eigenmaps across eleven sentence- and paragraph-level clustering tasks drawn from six dataset sources in the Massive Text Embedding Benchmark. Compared to standard k -NN graphs, our method outperforms in the low- k regime where disconnected components are prevalent, and matches standard k -NN at larger k . The advantage at low k persists when the standard k -NN graph is repaired with a minimum spanning tree, while our construction avoids the dense distance matrix an exact repair requires and is consequently far cheaper to build, in both time and memory. The code is available on https://github.com/bkolosk1/incremental_clustering_graphs.
- 15 May 2026
- Published by Elsevier in Information Sciences
- Vol. 754
Fully- and semi-supervised hierarchical multi-label image classification with graph learning
- 15 May 2026
- Published by Elsevier in Information Sciences
- Vol. 754
Fully- and semi-supervised hierarchical multi-label image classification with graph learning
- 1 April 2026
- Published by Springer Nature
- p. 360-372
Evaluation of Knowledge Graph Construction Methods on the Stroke Domain
- 1 April 2026
- Published by Springer Nature
- p. 360-372
Evaluation of Knowledge Graph Construction Methods on the Stroke Domain
- Open Access
- 17 March 2026
- Published by Royal Society of Chemistry (RSC) in Inorganic Chemistry Frontiers
- Vol. 13 (8) , 3213-3219
Digitalisation of inorganic chemistry with LLMs
Over the past few years, large language models have become technologically ubiquitous and now offer a powerful route to accelerate discoveries in inorganic chemistry.
- 17 March 2026
- Published by Royal Society of Chemistry (RSC) in Inorganic Chemistry Frontiers
- Vol. 13 (8) , 3213-3219
Open AccessDigitalisation of inorganic chemistry with LLMs
Over the past few years, large language models have become technologically ubiquitous and now offer a powerful route to accelerate discoveries in inorganic chemistry.
- Open Access
- 6 March 2026
- Published by Springer Nature in Machine Learning
- Vol. 115 (3) , 61
FuDoBa: Fusing Document and Knowledge Graph Based Representations with Bayesian Optimisation
Building on the success of large language models (LLMs), LLM-based representations have dominated the document representation landscape, achieving strong performance on document embedding benchmarks. However, high-dimensional, computationally expensive LLM embeddings can be too generic or inefficient for domain-specific and resource-scarce applications. To address these limitations, we introduce FuDoBa—a Bayesian optimisation-based representation learning method that integrates LLM embeddings with domain-specific structured knowledge, sourced both locally and from external repositories such as WikiData. This fusion produces low-dimensional, task-relevant representations while reducing training complexity and yielding interpretable early-fusion weights for improved classification performance. We demonstrate the effectiveness of our approach on six datasets across two domains, showing that when paired with robust AutoML-based classifiers, our method performs on par with, or surpasses, proprietary LLM-only embedding baselines, while offering modality-wise interpretability and a smaller dimensional footprint.
- 6 March 2026
- Published by Springer Nature in Machine Learning
- Vol. 115 (3) , 61
Open AccessFuDoBa: Fusing Document and Knowledge Graph Based Representations with Bayesian Optimisation
Building on the success of large language models (LLMs), LLM-based representations have dominated the document representation landscape, achieving strong performance on document embedding benchmarks. However, high-dimensional, computationally expensive LLM embeddings can be too generic or inefficient for domain-specific and resource-scarce applications. To address these limitations, we introduce FuDoBa—a Bayesian optimisation-based representation learning method that integrates LLM embeddings with domain-specific structured knowledge, sourced both locally and from external repositories such as WikiData. This fusion produces low-dimensional, task-relevant representations while reducing training complexity and yielding interpretable early-fusion weights for improved classification performance. We demonstrate the effectiveness of our approach on six datasets across two domains, showing that when paired with robust AutoML-based classifiers, our method performs on par with, or surpasses, proprietary LLM-only embedding baselines, while offering modality-wise interpretability and a smaller dimensional footprint.
- Open Access
PDK1: one abbreviation, two kinases, relentless confusion
The abbreviation PDK1 may refer to two different proteins: pyruvate dehydrogenase kinase 1 and 3-phosphoinositide-dependent protein kinase 1. This overlap introduces ambiguity, making it challenging to discern which protein is being referenced. Here, we highlight widespread confusion surrounding PDK1, including cases where articles have stated incorrect antibodies, referred to incorrect sequences for PCR, gene silencing, or plasmid construction, merged the properties of pyruvate dehydrogenase kinase 1 and 3-phosphoinositide-dependent protein kinase 1, or incorrectly cited the other protein. Notably, 19 % of articles with the term “PDK1” on PubMed, published between 2019 and the middle of 2025, contain at least one such mistake. This ambiguity extends beyond the scientific literature to websites of biotechnology providers and vendors, where antibodies or recombinant proteins are misattributed. To mitigate this issue, the use of unique protein abbreviations, clear antibody and sequence identification, and a more rigorous peer review process are needed.
Open AccessPDK1: one abbreviation, two kinases, relentless confusion
The abbreviation PDK1 may refer to two different proteins: pyruvate dehydrogenase kinase 1 and 3-phosphoinositide-dependent protein kinase 1. This overlap introduces ambiguity, making it challenging to discern which protein is being referenced. Here, we highlight widespread confusion surrounding PDK1, including cases where articles have stated incorrect antibodies, referred to incorrect sequences for PCR, gene silencing, or plasmid construction, merged the properties of pyruvate dehydrogenase kinase 1 and 3-phosphoinositide-dependent protein kinase 1, or incorrectly cited the other protein. Notably, 19 % of articles with the term “PDK1” on PubMed, published between 2019 and the middle of 2025, contain at least one such mistake. This ambiguity extends beyond the scientific literature to websites of biotechnology providers and vendors, where antibodies or recombinant proteins are misattributed. To mitigate this issue, the use of unique protein abbreviations, clear antibody and sequence identification, and a more rigorous peer review process are needed.
- Open Access
- 11 November 2025
- Published by MDPI AG in Machine Learning and Knowledge Extraction
- Vol. 7 (4) , 142
Extreme Multi-Label Text Classification for Less-Represented Languages and Low-Resource Environments: Advances and Lessons Learned
Amid ongoing efforts to develop extremely large, multimodal models, there is increasing interest in efficient Small Language Models (SLMs) that can operate without reliance on large data-centre infrastructure. However, recent SLMs (e.g., LLaMA or Phi) with up to three billion parameters are predominantly trained in high-resource languages, such as English, which limits their applicability to industries that require robust NLP solutions for less-represented languages and low-resource settings, particularly those requiring low latency and adaptability to evolving label spaces. This paper examines a retrieval-based approach to multi-label text classification (MLC) for a media monitoring dataset, with a particular focus on less-represented languages, such as Slovene. This dataset presents an extreme MLC challenge, with instances labelled using up to twelve thousand categories. The proposed method, which combines retrieval with computationally efficient prediction, effectively addresses challenges related to multilinguality, resource constraints, and frequent label changes. We adopt a model-agnostic approach that does not rely on a specific model architecture or language selection. Our results demonstrate that techniques from the extreme multi-label text classification (XMC) domain outperform traditional Transformer-based encoder models, particularly in handling dynamic label spaces without requiring continuous fine-tuning. Additionally, we highlight the effectiveness of this approach in scenarios involving rare labels, where baseline models struggle with generalisation.
- 11 November 2025
- Published by MDPI AG in Machine Learning and Knowledge Extraction
- Vol. 7 (4) , 142
Open AccessExtreme Multi-Label Text Classification for Less-Represented Languages and Low-Resource Environments: Advances and Lessons Learned
Amid ongoing efforts to develop extremely large, multimodal models, there is increasing interest in efficient Small Language Models (SLMs) that can operate without reliance on large data-centre infrastructure. However, recent SLMs (e.g., LLaMA or Phi) with up to three billion parameters are predominantly trained in high-resource languages, such as English, which limits their applicability to industries that require robust NLP solutions for less-represented languages and low-resource settings, particularly those requiring low latency and adaptability to evolving label spaces. This paper examines a retrieval-based approach to multi-label text classification (MLC) for a media monitoring dataset, with a particular focus on less-represented languages, such as Slovene. This dataset presents an extreme MLC challenge, with instances labelled using up to twelve thousand categories. The proposed method, which combines retrieval with computationally efficient prediction, effectively addresses challenges related to multilinguality, resource constraints, and frequent label changes. We adopt a model-agnostic approach that does not rely on a specific model architecture or language selection. Our results demonstrate that techniques from the extreme multi-label text classification (XMC) domain outperform traditional Transformer-based encoder models, particularly in handling dynamic label spaces without requiring continuous fine-tuning. Additionally, we highlight the effectiveness of this approach in scenarios involving rare labels, where baseline models struggle with generalisation.
- 23 September 2025
- Published by Springer Nature
- p. 181-197
From Symbolic to Neural and Back: Exploring Knowledge Graph–Large Language Model Synergies
- 23 September 2025
- Published by Springer Nature
- p. 181-197
From Symbolic to Neural and Back: Exploring Knowledge Graph–Large Language Model Synergies
- Open Access
FuDoBa: Fusing Document and Knowledge Graph-based Representations with Bayesian Optimisation
Building on the success of Large Language Models (LLMs), LLM-based representations have dominated the document representation landscape, achieving great performance on the document embedding benchmarks. However, the high-dimensional, computationally expensive embeddings from LLMs tend to be either too generic or inefficient for domain-specific applications. To address these limitations, we introduce FuDoBa a Bayesian optimisation-based method that integrates LLM-based embeddings with domain-specific structured knowledge, sourced both locally and from external repositories like WikiData. This fusion produces low-dimensional, task-relevant representations while reducing training complexity and yielding interpretable early-fusion weights for enhanced classification performance. We demonstrate the effectiveness of our approach on six datasets in two domains, showing that when paired with robust AutoML-based classifiers, our proposed representation learning approach performs on par with, or surpasses, those produced solely by the proprietary LLM-based embedding baselines.
Open AccessFuDoBa: Fusing Document and Knowledge Graph-based Representations with Bayesian Optimisation
Building on the success of Large Language Models (LLMs), LLM-based representations have dominated the document representation landscape, achieving great performance on the document embedding benchmarks. However, the high-dimensional, computationally expensive embeddings from LLMs tend to be either too generic or inefficient for domain-specific applications. To address these limitations, we introduce FuDoBa a Bayesian optimisation-based method that integrates LLM-based embeddings with domain-specific structured knowledge, sourced both locally and from external repositories like WikiData. This fusion produces low-dimensional, task-relevant representations while reducing training complexity and yielding interpretable early-fusion weights for enhanced classification performance. We demonstrate the effectiveness of our approach on six datasets in two domains, showing that when paired with robust AutoML-based classifiers, our proposed representation learning approach performs on par with, or surpasses, those produced solely by the proprietary LLM-based embedding baselines.
- 2 May 2025
- Published by Springer Nature
- p. 261-273
Make Literature-Based Discovery Great Again Through Reproducible Pipelines
- 2 May 2025
- Published by Springer Nature
- p. 261-273
Make Literature-Based Discovery Great Again Through Reproducible Pipelines