Dolby (Netherlands) researcher spanning multimodal vision-language reasoning and video-aligned generative music
2
papers here
1
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
0
h-index · OpenAlex
Bibliometric figures come from OpenAlex (2026-10-17) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.
Research profile
Output—
This axis measures total publication volume and venue record. The current snapshot cannot score it: only two 2026 arXiv preprints are captured, with no consolidated works list, so output volume cannot be reliably quantified yet.
Impact—
This axis tracks citation-based influence such as h-index and i10-index. It is unscored here: the snapshot reports h-index 0 and i10-index 0, which reflects an unresolved or too-new citation footprint rather than a measured impact level.
Focus100/100
Scored 100, the maximum. The snapshot's field distribution lists a single labeled field — machine learning (multimodal vision-language models) — and the three recorded topics (generative adversarial networks and image synthesis; music and audio processing; music technology and sound studies) each carry a count of 2, clustering into two coherent threads: multimodal visual reasoning and generative audio/music. This is a tightly concentrated, dual-track research profile with no scattered outlier topics.
Collaboration58/100
Scored 58, a mid-range collaborative footprint built on 17 distinct coauthors in the snapshot. Twelve recurring collaborators are listed — including Sreyan Ghosh, Ramani Duraiswami, Dinesh Manocha, Sunli Chen, Ding Zhong, Ziqiao Ma, Jiaxin Liu, and Zeyuan Yang — and both catalog papers are middle-author team efforts with no corresponding-author role, indicating contribution within larger multi-author groups rather than solo or led work.
Activity32/100
Scored 32, a low-to-moderate activity level. The snapshot's countsByYear records exactly 2 works in 2026, matching recent3yWorks of 2 — meaning all captured output falls in a single year and consists of arXiv preprints. This thin, recent footprint keeps the activity score low; it likely reflects an emerging or newly resolved OpenAlex identity rather than a long publication record.
Reproducibility55/100
Scored 55, computed from a mean reproducibility rating of 55 across just 1 scored paper. The scored paper, 'Multimodal Thinking with Renderable Programs', is rated medium for reproducibility readiness: Evolve's verdict credits a reported ~6.5-10 weighted-point gain over SFT on geometry, but notes that no significance tests and no code or model release are stated — leaving empirical claims partly unverifiable.
Research summary
Lie Lu is a researcher at Dolby (Netherlands) whose indexed work spans two adjacent threads of generative AI: multimodal vision-language reasoning and generative audio for video. This profile is name-resolved from public OpenAlex bibliographic data, as no ORCID is linked. The catalog contains two 2026 arXiv preprints, both with Lu in a middle author position: 'Multimodal Thinking with Renderable Programs', which Evolve has reviewed, and 'VIBE: Video Instruction-aligned Background music gEneration', which has not yet been reviewed. In the reviewed paper, the SVGLM method tunes vision-language models to sketch SVG overlays while reasoning, reportedly gaining roughly 6.5-10 weighted points over SFT on geometry tasks; Evolve notes the absence of stated significance tests or a release, and rates reproducibility readiness as medium. The snapshot's topic profile — generative adversarial networks and image synthesis, music and audio processing, and music technology and sound studies, each with a count of 2 — reinforces this dual focus. Collaboration is moderate: 17 distinct coauthors and 12 recurring collaborators, with neither catalog paper led or corresponding-authored. The activity footprint is thin — 2 works recorded, both in 2026 — and the impact axis is unscored, with h-index and i10-index reported as 0. Overall, this is an early-visibility, team-based industrial research profile concentrated on making models reason visually and generate audio in step with video content.
Collaboration pattern
Lie Lu works within multi-author teams: 17 distinct coauthors appear in the snapshot, with 12 recurring collaborators listed, spanning both the audio thread (e.g., Sreyan Ghosh, Ramani Duraiswami, Dinesh Manocha) and the vision-language thread (e.g., Sunli Chen, Ding Zhong, Ziqiao Ma, Zeyuan Yang). Both catalog papers place Lu in a middle author position with no corresponding role, consistent with the mid-range collaboration score of 58 — contributing within groups rather than leading them.
Research keywords
Highlights
SVG overlays as a visual chain-of-thought medium (SVGLM)
Video instruction-aligned background music generation (VIBE)
Co-authored 'VIBE: Video Instruction-aligned Background music gEneration' (arXiv 2026, middle author), targeting background music generation aligned with video instructions; this preprint sits in the catalog without an Evolve verdict, keywords, or reproducibility rating yet.
A tightly focused dual-track topic profile
The bibliometric snapshot records three topic clusters — generative adversarial networks and image synthesis, music and audio processing, and music technology and sound studies — each with a count of 2, alongside a maximum focus score of 100, indicating concentrated work on visual reasoning and generative audio rather than scattered interests.
Timeline
- 2026Co-authored 'Multimodal Thinking with Renderable Programs' (arXiv), middle author position; the paper is Evolve-reviewed with medium reproducibility readiness.
- 2026Co-authored 'VIBE: Video Instruction-aligned Background music gEneration' (arXiv), middle author position; not yet reviewed by Evolve.
- 2026OpenAlex snapshot records 2 indexed works for the year (recent3yWorks = 2), with 17 distinct coauthors across the profile.
Papers in the Evolve catalog
Frequent co-authors
About this profile
This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.
- Bibliometrics come from an OpenAlex identity snapshot only; no ORCID is linked, and 'Lie Lu' is a name shared across authors, so this profile is name-resolved from public bibliographic data.
- Output and impact axes are unscored: the snapshot reports h-index 0 and i10-index 0 with only 2 indexed works, so citation influence cannot yet be assessed.
- Only one of the two catalog papers has an Evolve verdict; the VIBE preprint lacks reviewed keywords, a reproducibility rating, and a catalog slug.
- All captured works are 2026 arXiv preprints; no peer-reviewed venue, funding, or award information is present in this data scope.