AI Level of Detail: Distance-Aware ML Model Precision Selection for Real-Time Human Motion Prediction in Games

First page of the AI Level of Detail research paper

AI Level of Detail (AI LOD) extends the classic geometry level-of-detail idea to machine learning inference. Just as a game engine swaps a distant mesh for a cheaper one when the difference is imperceptible, AI LOD swaps a full-precision motion model for a cheaper quantized one based on how far an NPC is from the player camera. The novelty is the concept itself: inference-time quantization precision as the LOD axis for AI-driven character animation.

Using the ConvSeq2Seq human-motion predictor of Li et al. as a fixed backbone, the trained checkpoint is exported into three ONNX Runtime tiers — FP32 (closest NPCs), FP16 (mid-range), and INT8 per-tensor (farthest) — routed at runtime by a distance-based selector. The tiers are evaluated on the CMU Mocap dataset for latency, model size, and motion accuracy, alongside a small perceptual study.

Key results (CPU, vs FP32 baseline):

  • FP16 — 1.53× faster, 49% smaller, near-zero accuracy loss (relative L2 7.8e-4).
  • INT8 per-tensor — 9.79× faster, 73% smaller, with degradation that stays below the perceptual threshold at far distances.

A 15-participant perceptual study found no detectable difference when each tier was viewed within its intended distance range, supporting the core premise that distance-aware precision selection is a viable LOD strategy for AI-based animation.

Venue

Presented at the PRICE Workshop (Performance, Rendering & Interaction in Competitive Esports), SIGGRAPH 2026 Technical Workshops — Los Angeles Convention Center, July 19–23, 2026. Talk given in Technical Sketch Session 2 on Thursday, July 23, 2026.

Photos

Mathew Varghese presenting the AI Level of Detail talk at a podium beside a projected SIGGRAPH slide
Presenting AI Level of Detail in Technical Sketch Session 2, PRICE Workshop, SIGGRAPH 2026.
Three panelists seated at a table with microphones during the PRICE Workshop panel discussion
Panel discussion following the technical sketch session, PRICE Workshop, SIGGRAPH 2026.
Diagram showing model precision tiers selected based on distance from the camera
Distance-aware model precision selection: FP32, FP16, and INT8 tiers routed by NPC distance from the camera.

Links

Resource Link
Paper (ACM Digital Library) dl.acm.org/doi/10.1145/3799828.3816004
Preprint (arXiv) arXiv:2606.06565
PDF Download (hosted copy)
Presentation slides Google Slides deck
Workshop schedule price-workshop.org/2026/schedule/
LinkedIn post linkedin.com/posts/mathewvarghesemanu

Citation

Mathew Varghese. 2026. AI Level of Detail: Distance-Aware ML Model Precision Selection for Real-Time Human Motion Prediction in Games. In Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Technical Workshops (SIGGRAPH Technical Workshops ’26). ACM, New York, NY, USA, 1–4. https://doi.org/10.1145/3799828.3816004