--- title: "Ivana Balažević — author profile and papers" description: "Ivana Balažević, Google DeepMind: 2 papers on Elamur Research (2019–2022), h-index 9 — affiliations, topics, co-authors and who cites them in our library." canonical: "https://research.elamur.ai/free/en/authors/ivana-balazevic/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/ivana-balazevic/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Ivana Balažević # Ivana Balažević Ivana Balažević works on machine learning for language and structured knowledge. Their papers here follow two lines of work. The first is link prediction in knowledge graphs, where missing subject–relation–object facts are inferred from those already stored. The second is few-shot learning with pretrained masked language models. Both lines aim at compact models: factorizations whose size stays modest as graphs grow, and ways of adapting language models that need little prompt engineering and few task-specific parameters. The most cited of these works is [TuckER](https://md.research.elamur.ai/free/en/papers/balazevic-2019-tucker-tensor-factorization-knowledge-graph/) (2019). It treats a knowledge graph as a three-way binary tensor and completes it with Tucker decomposition. Rows of the factor matrices act as embeddings for entities and relations. A small core tensor, shared by every relation, governs how those embeddings interact. This lets relations pass information to one another, and the parameter count scales linearly as entities and relations are added. The more recent paper, [Cutting Down on Prompts and Parameters](https://md.research.elamur.ai/free/en/papers/iv-2022-cutting-down-prompts-parameters-simple/) (2022), moves to language models. It asks whether few-shot classifiers can be easy to prompt and also memory-efficient. The study uses RoBERTa and ALBERT on eight GLUE and SuperGLUE tasks, with sixteen labelled examples per class. It proposes "null prompts", which join the text fields of each example to a mask token without template wording or demonstrations. Under finetuning, these prompts come close to manually designed ones. Balažević is currently at [Google DeepMind](https://md.research.elamur.ai/free/en/companies/google-deepmind/) in London. Earlier bylines list the [University of Edinburgh](https://md.research.elamur.ai/free/en/universities/university-of-edinburgh/) (2018–2021) and [Meta](https://md.research.elamur.ai/free/en/companies/meta/) in Menlo Park (2019–2020). Co-authors on the papers here include [Eric Wallace](https://md.research.elamur.ai/free/en/authors/eric-wallace/), [Sameer Singh](https://md.research.elamur.ai/free/en/authors/sameer-singh/), [Fabio Petroni](https://md.research.elamur.ai/free/en/authors/fabio-petroni/) and [Sebastian Riedel](https://md.research.elamur.ai/free/en/authors/sebastian-riedel/). The work falls under [Knowledge graphs & ontologies](https://md.research.elamur.ai/free/en/topics/knowledge-graphs/) and [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/). Ivana Balažević has 2 papers in our library, in 2019–2022, most of them on [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/), [Knowledge graphs & ontologies](https://md.research.elamur.ai/free/en/topics/knowledge-graphs/); often with [Eric Wallace](https://md.research.elamur.ai/free/en/authors/eric-wallace/), [Fabio Petroni](https://md.research.elamur.ai/free/en/authors/fabio-petroni/), [Sameer Singh](https://md.research.elamur.ai/free/en/authors/sameer-singh/). 2 papers in our library · 2019–2022 [OpenAlex](https://openalex.org/A5035972465) · [ORCID](https://orcid.org/0000-0001-8268-1484) · [**h-index 9**](https://openalex.org/authors/A5035972465) · [21 works](https://openalex.org/works?filter=author.id:A5035972465) · [1,171 citations](https://openalex.org/works?filter=author.id:A5035972465&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [📄 2 works in our library](https://md.research.elamur.ai/free/en/authors/ivana-balazevic/#papers) · [🔗 334 citations in our library](https://md.research.elamur.ai/free/en/authors/ivana-balazevic/cited-by/) · [1 more in our library without a page yet](#also) · [⭐ Paper Star rank #1982](https://md.research.elamur.ai/free/en/paper-stars/page/20/#star-1982) The profile is the library's match of the name; two people with one name are not always told apart. Elsewhere: [Google Scholar](https://scholar.google.com/citations?view_op=search_authors&mauthors=Ivana+Bala%C5%BEevi%C4%87) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Ivana+Bala%C5%BEevi%C4%87) · [dblp](https://dblp.org/search?q=Ivana+Bala%C5%BEevi%C4%87) · [ACL Anthology](https://aclanthology.org/search/?q=Ivana+Bala%C5%BEevi%C4%87) ## Affiliations [🏢**Google DeepMind** London · United Kingdom 2022–2025 · 2 papers*current*](https://md.research.elamur.ai/free/en/companies/google-deepmind/) · [🎓**University of Edinburgh** Edinburgh · United Kingdom 2018–2021 · 7 papers](https://md.research.elamur.ai/free/en/universities/university-of-edinburgh/) · [🏢**Meta** Menlo Park · United States 2019–2020 · 2 papers](https://md.research.elamur.ai/free/en/companies/meta/) From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. ## Topics [Language models & pretraining (1)](https://md.research.elamur.ai/free/en/topics/language-models/) · [Knowledge graphs & ontologies (1)](https://md.research.elamur.ai/free/en/topics/knowledge-graphs/) ## Works by year ### 2020s [**2022** 1 paper](https://md.research.elamur.ai/free/en/papers/iv-2022-cutting-down-prompts-parameters-simple/) ### 2010s [**2019** 1 paper](https://md.research.elamur.ai/free/en/papers/balazevic-2019-tucker-tensor-factorization-knowledge-graph/) ## Papers The most cited first (OpenAlex and our library). 1. **[TuckER: Tensor Factorization for Knowledge Graph Completion](https://md.research.elamur.ai/free/en/papers/balazevic-2019-tucker-tensor-factorization-knowledge-graph/)** *(2019 · 792 citations)* Knowledge graphs store facts as "subject – relation – object" triples but are almost always incomplete. Methods are therefore needed that recover missing links, i.e., that solve the link prediction task. The authors propose TuckER, a linear model. It represents the graph as a three-dimensional binary tensor and… 2. **[Cutting Down on Prompts and Parameters: Simple Few-Shot Learning with Language Models](https://md.research.elamur.ai/free/en/papers/iv-2022-cutting-down-prompts-parameters-simple/)** *(2022 · 215 citations)* Few-shot learning with masked language models usually depends on carefully crafted prompts. Prompt-based finetuning is accurate but needs a full model copy for every task, while in-context learning reuses one frozen model but is very sensitive to prompt wording. The authors test whether simple prompt design and… ## Also in our library 1 more work of theirs is in our library without a page here yet; it will appear as we publish them. ## Co-authors here 4 co-authors with a page, the most papers together first. - [Eric Wallace](https://md.research.elamur.ai/free/en/authors/eric-wallace/) - [Fabio Petroni](https://md.research.elamur.ai/free/en/authors/fabio-petroni/) - [Sameer Singh](https://md.research.elamur.ai/free/en/authors/sameer-singh/) - [Sebastian Riedel](https://md.research.elamur.ai/free/en/authors/sebastian-riedel/) ## Cited by Papers of our library cite their works 334 times (of the 140,884 papers whose bibliographies we read); [25 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/ivana-balazevic/cited-by/).