--- title: "Linjie Li — author profile and papers" description: "Linjie Li, Microsoft (United States): 4 papers on Elamur Research (2019–2023), h-index 28 — affiliations, topics, co-authors and who cites them in our…" canonical: "https://research.elamur.ai/free/en/authors/linjie-li/" --- [Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/authors/linjie-li/) [home](https://md.research.elamur.ai/free/en/) / [authors](https://md.research.elamur.ai/free/en/authors/) / Linjie Li # Linjie Li Linjie Li works on vision and language, building and testing models that read images and text together. Li's papers here move from learning shared image-text representations through large-scale pre-training to the evaluation and reliability of large multimodal models (LMMs). Recurring questions include how one model can serve many vision-language tasks, which combined skills such models actually have, and why they sometimes describe content absent from an image. The methods range from Transformer pre-training on public image-caption corpora to benchmark design, hand-built qualitative test sets, and instruction tuning on generated data. The most cited work is [UNITER](https://md.research.elamur.ai/free/en/papers/chen-2019-uniter-universal-image-text-representation/) (2019). It trains a single Transformer on region features and word tokens to yield a general embedding that transfers across many image-text tasks. Its pre-training corrupts only one modality at a time, so the model always sees the other one intact. Among the 2023 papers, [MM-Vet](https://md.research.elamur.ai/free/en/papers/yu-2023-mm-vet-evaluating-large-multimodal/) is a benchmark that treats hard tasks, such as explaining a meme, as blends of six basic abilities, including OCR, math, spatial awareness and world knowledge. Because answers can be anything from one word to a long essay, GPT-4 grades the responses instead of exact matching. The newest work turns to large multimodal models. [The Dawn of LMMs](https://md.research.elamur.ai/free/en/papers/yang-2023-dawn-lmms-preliminary-explorations-gpt/) (2023) probes GPT-4V(ision) qualitatively with curated cases spanning documents, charts, video, puzzles and code generation from images, and it reports no benchmark scores. [Mitigating Hallucination in Large Multi-Modal Models](https://md.research.elamur.ai/free/en/papers/liu-2023-mitigating-hallucination-large-multi-modal/) (2023) traces invented content to a reliance on language priors and to training sets made almost entirely of positive examples. As a fix, it introduces LRV-Instruction, a 400k-example corpus that pairs ordinary prompts with deliberately misleading ones. Li is affiliated with [Microsoft](https://md.research.elamur.ai/free/en/companies/microsoft/) in Redmond, with earlier papers from the [University of California San Diego](https://md.research.elamur.ai/free/en/universities/university-of-california-san-diego/). Frequent co-authors include [Jianfeng Wang](https://md.research.elamur.ai/free/en/authors/jianfeng-wang/), [Kevin Lin](https://md.research.elamur.ai/free/en/authors/kevin-lin/) and [Lijuan Wang](https://md.research.elamur.ai/free/en/authors/lijuan-wang/), along with Zhengyuan Yang and Zicheng Liu. The work sits mainly in [Vision & language](https://md.research.elamur.ai/free/en/topics/multimodal/), with ties to language models and pre-training. 4 papers on this site · 2019–2023 current affiliation [Microsoft (United States)](https://md.research.elamur.ai/free/en/companies/microsoft/), Redmond, United States; [University of Electronic Science and Technology of China](https://md.research.elamur.ai/free/en/universities/university-of-electronic-science-and-technology-of-china/), Chengdu, China Linjie Li has 4 papers on this site, in 2019–2023, most of them on [Vision & language](https://md.research.elamur.ai/free/en/topics/multimodal/), [Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/); often with [Jianfeng Wang](https://md.research.elamur.ai/free/en/authors/jianfeng-wang/), [Kevin Lin](https://md.research.elamur.ai/free/en/authors/kevin-lin/), [Lijuan Wang](https://md.research.elamur.ai/free/en/authors/lijuan-wang/). [OpenAlex](https://openalex.org/A5100657555) · [ORCID](https://orcid.org/0000-0003-0867-8863) · [**h-index 28**](https://openalex.org/authors/A5100657555) · [97 works](https://openalex.org/works?filter=author.id:A5100657555) · [6 331 citations](https://openalex.org/works?filter=author.id:A5100657555&sort=cited_by_count:desc) OpenAlex, 2026-10-14 [4 works on this site](https://md.research.elamur.ai/free/en/authors/linjie-li/#papers) · [1 230 citations in our library](https://md.research.elamur.ai/free/en/authors/linjie-li/cited-by/) · 3 more works in our library · [⭐ Paper Star rank #816](https://md.research.elamur.ai/free/en/paper-stars/page/9/#star-816) 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=Linjie+Li) · [Semantic Scholar](https://www.semanticscholar.org/search?q=Linjie+Li) · [dblp](https://dblp.org/search?q=Linjie+Li) · [ACL Anthology](https://aclanthology.org/search/?q=Linjie+Li) ## Affiliations - [Microsoft (United States)](https://md.research.elamur.ai/free/en/companies/microsoft/) · Redmond, United States —current · 2019–2026 on 43 papers - [University of Electronic Science and Technology of China](https://md.research.elamur.ai/free/en/universities/university-of-electronic-science-and-technology-of-china/) · Chengdu, China —current · 1996–2026 on 4 papers - [Beijing University of Posts and Telecommunications](https://md.research.elamur.ai/free/en/universities/beijing-university-of-posts-and-telecommunications/) · Beijing, China —2024–2025 on 2 papers - [Beijing University of Technology](https://ror.org/037b1pp87) · Beijing, China —2023–2025 on 2 papers - [Carnegie Mellon University](https://md.research.elamur.ai/free/en/universities/carnegie-mellon-university/) · Pittsburgh, United States —2024–2025 on 2 papers - [University of California San Diego](https://md.research.elamur.ai/free/en/universities/university-of-california-san-diego/) · San Diego, United States —2016–2021 on 4 papers From the bylines of their papers (OpenAlex, 2026-10-14), ORCID and Wikidata; institutions on papers are matched by machine. 9 more institutions on single bylines are not listed. ## Topics [Vision & language 4](https://md.research.elamur.ai/free/en/topics/multimodal/) · [Language models & pretraining 2](https://md.research.elamur.ai/free/en/topics/language-models/) ## Works by year ### 2020s [**2023** 3 papers](https://md.research.elamur.ai/free/en/authors/linjie-li/year/2023/) ### 2010s [**2019** 1 paper](https://md.research.elamur.ai/free/en/papers/chen-2019-uniter-universal-image-text-representation/) ## Papers The most cited first (OpenAlex and our library). 1. **[UNITER: UNiversal Image-TExt Representation Learning](https://md.research.elamur.ai/free/en/papers/chen-2019-uniter-universal-image-text-representation/)** *(2019 · 2 344 citations)* Most Vision-and-Language (V+L) models build joint image-text representations designed for one task, so they rarely transfer to other tasks. 2. **[MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilities](https://md.research.elamur.ai/free/en/papers/yu-2023-mm-vet-evaluating-large-multimodal/)** *(2023 · 328 citations)* Classic vision-language benchmarks such as VQA v2, COCO captioning, and TextVQA each probe only one or two skills. 3. **[The Dawn of LMMs: Preliminary Explorations with GPT-4V(ision)](https://md.research.elamur.ai/free/en/papers/yang-2023-dawn-lmms-preliminary-explorations-gpt/)** *(2023 · 303 citations)* Microsoft researchers present an early, mostly qualitative study of GPT-4V(ision), the vision-enabled version of GPT-4 from OpenAI. 4. **[Mitigating Hallucination in Large Multi-Modal Models via Robust Instruction Tuning](https://md.research.elamur.ai/free/en/papers/liu-2023-mitigating-hallucination-large-multi-modal/)** *(2023 · 180 citations)* Large multi-modal models (LMMs) such as MiniGPT4, LLaVA, InstructBLIP, mPLUG-Owl and MMGPT often describe things that are not in the image, or they ignore the user's request and return a generic caption. ## Also in our library 3 more works of theirs are in our library without a page here yet; they will appear as we publish them. ## Co-authors here 12 of 16 co-authors with a page, the most papers together first. - [Jianfeng Wang](https://md.research.elamur.ai/free/en/authors/jianfeng-wang/) ×3 - [Kevin Lin](https://md.research.elamur.ai/free/en/authors/kevin-lin/) ×3 - [Lijuan Wang](https://md.research.elamur.ai/free/en/authors/lijuan-wang/) ×3 - [Zhengyuan Yang](https://md.research.elamur.ai/free/en/authors/zhengyuan-yang/) ×2 - [Zicheng Liu](https://md.research.elamur.ai/free/en/authors/zicheng-liu/) ×2 - [Ahmed El Kholy](https://md.research.elamur.ai/free/en/authors/ahmed-el-kholy/) - [Faisal Ahmed](https://md.research.elamur.ai/free/en/authors/faisal-ahmed/) - [Fuxiao Liu](https://md.research.elamur.ai/free/en/authors/fuxiao-liu/) - [Jingjing Liu](https://md.research.elamur.ai/free/en/authors/jingjing-liu/) - [Licheng Yu](https://md.research.elamur.ai/free/en/authors/licheng-yu/) - [Weihao Yu](https://md.research.elamur.ai/free/en/authors/weihao-yu/) - [Xinchao Wang](https://md.research.elamur.ai/free/en/authors/xinchao-wang/) ## Cited by Papers of our library cite their works 1 230 times (of the 140 884 papers whose bibliographies we read); [98 citing papers with a page here →](https://md.research.elamur.ai/free/en/authors/linjie-li/cited-by/).