--- title: "Felix Gimeno — author profile and papers" description: "Felix Gimeno: 2 papers on Elamur Research (2021–2022), h-index 7 — affiliations, topics, co-authors and who cites them in our library." canonical: "https://research.elamur.ai/free/en/authors/felix-gimeno/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/felix-gimeno/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Felix Gimeno # Felix Gimeno Felix Gimeno's papers here, published in 2021 and 2022, apply large Transformer-based neural networks to two quite different problems. One is writing complete, working programs for competitive programming tasks from long natural-language problem statements. The other is designing a single perception architecture that can take in images, audio and other inputs without being rebuilt for each modality. Both works deal with making attention models cope with long or very large inputs, and both put weight on evaluation that is hard to game. The more cited of the two is [Competition-Level Code Generation with AlphaCode](https://md.research.elamur.ai/free/en/papers/li-2022-competition-level-code-generation-alphacode/) (2022). It describes a DeepMind system that pre-trains Transformers on curated GitHub code and then fine-tunes them on CodeContests, a newly released dataset of contest problems. The dataset is divided by publication date so that test problems postdate the training material, and extra generated tests make it harder for incorrect programs to pass. The system samples a very large number of candidate solutions for each problem. The earlier [Perceiver: General Perception with Iterative Attention](https://md.research.elamur.ai/free/en/papers/jaegle-2021-perceiver-general-perception-iterative-attention/) (2021) addresses the quadratic cost of standard attention. A compact learned latent array repeatedly cross-attends to the raw input, and a deep self-attention stack then works on the latents alone. As a result, compute grows linearly with input size and network depth no longer depends on input length. The move from the general perception model of 2021 to the code generation work of 2022 marks a turn toward [Code & programming languages](https://md.research.elamur.ai/free/en/topics/code-and-programming/), with language models asked to choose algorithms and produce full solutions rather than short snippets. Paper bylines from 2019 to 2020 list [Google](https://md.research.elamur.ai/free/en/companies/google/) in Mountain View, United States. Co-authors on the papers here include [Oriol Vinyals](https://md.research.elamur.ai/free/en/authors/oriol-vinyals/), [Andrew Zisserman](https://md.research.elamur.ai/free/en/authors/andrew-zisserman/), [Junyoung Chung](https://md.research.elamur.ai/free/en/authors/junyoung-chung/) and [Nate Kushman](https://md.research.elamur.ai/free/en/authors/nate-kushman/). Felix Gimeno has 2 papers in our library, in 2021–2022, most of them on [Code & programming languages](https://md.research.elamur.ai/free/en/topics/code-and-programming/); often with [Andrew Zisserman](https://md.research.elamur.ai/free/en/authors/andrew-zisserman/), [Junyoung Chung](https://md.research.elamur.ai/free/en/authors/junyoung-chung/), [Nate Kushman](https://md.research.elamur.ai/free/en/authors/nate-kushman/). 2 papers in our library · 2021–2022 [OpenAlex](https://openalex.org/A5000173050) · [ORCID](https://orcid.org/0000-0002-1105-048X) · [**h-index 7**](https://openalex.org/authors/A5000173050) · [9 works](https://openalex.org/works?filter=author.id:A5000173050) · [1,282 citations](https://openalex.org/works?filter=author.id:A5000173050&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [📄 2 works in our library](https://md.research.elamur.ai/free/en/authors/felix-gimeno/#papers) · [🔗 647 citations in our library](https://md.research.elamur.ai/free/en/authors/felix-gimeno/cited-by/) · [⭐ Paper Star rank #1301](https://md.research.elamur.ai/free/en/paper-stars/page/14/#star-1301) 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=Felix+Gimeno) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Felix+Gimeno) · [dblp](https://dblp.org/search?q=Felix+Gimeno) · [ACL Anthology](https://aclanthology.org/search/?q=Felix+Gimeno) ## Affiliations [🏢**Google** Mountain View · United States 2019–2020 · 3 papers](https://md.research.elamur.ai/free/en/companies/google/) From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 1 more institution on single bylines are not listed. ## Topics [Code & programming languages (1)](https://md.research.elamur.ai/free/en/topics/code-and-programming/) ## Works by year ### 2020s [**2022** 1 paper](https://md.research.elamur.ai/free/en/papers/li-2022-competition-level-code-generation-alphacode/) · [**2021** 1 paper](https://md.research.elamur.ai/free/en/papers/jaegle-2021-perceiver-general-perception-iterative-attention/) ## Papers The most cited first (OpenAlex and our library). 1. **[Competition-Level Code Generation with AlphaCode](https://md.research.elamur.ai/free/en/papers/li-2022-competition-level-code-generation-alphacode/)** *(2022 · 1,543 citations)* Large language models can already write short code snippets. Full competitive programming problems are much harder: the model has to read a long problem statement, choose an algorithm and write a complete program that passes hidden tests. AlphaCode, from DeepMind, tackles this setting. Transformer models are… 2. **[Perceiver: General Perception with Iterative Attention](https://md.research.elamur.ai/free/en/papers/jaegle-2021-perceiver-general-perception-iterative-attention/)** *(2021 · 238 citations)* Most deep perception models are tailored to a single modality. ConvNets, for example, rely on the 2D grid structure of images, so each new input type or sensor combination needs a redesigned architecture. Transformers make fewer assumptions about their inputs, but their cost grows quadratically with input length… ## Co-authors here 4 co-authors with a page, the most papers together first. - [Andrew Zisserman](https://md.research.elamur.ai/free/en/authors/andrew-zisserman/) - [Junyoung Chung](https://md.research.elamur.ai/free/en/authors/junyoung-chung/) - [Nate Kushman](https://md.research.elamur.ai/free/en/authors/nate-kushman/) - [Oriol Vinyals](https://md.research.elamur.ai/free/en/authors/oriol-vinyals/) ## Cited by Papers of our library cite their works 647 times (of the 140,884 papers whose bibliographies we read); [57 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/felix-gimeno/cited-by/).