Liquid Neural Networks Improve Dynamic 3D Scene Reconstruction from Video
Researchers have proposed replacing the standard MLP deformation field in Deformable 3D Gaussian Splatting with Closed-form Continuous-time (CfC) cells, a type of Liquid Neural Network, to better handle temporal dynamics in dynamic scene reconstruction. The existing MLP-based approach, while fitted to a continuous time variable, effectively produces discrete per-frame offsets with temporal smoothness only emerging incidentally during optimization. The proposed architecture bakes continuous-time behavior directly into the model, yielding improved performance particularly on scenes with high-frequency articulated motion.
A preprint submitted to arXiv introduces a modification to Deformable 3D Gaussian Splatting (D-3DGS), a technique for reconstructing dynamic scenes from monocular video. The authors argue that the conventional positional-encoded MLP used as a deformation field does not structurally couple different time values, making it functionally discrete despite operating on a continuous time variable. To address this, they replace the MLP with a stack of Closed-form Continuous-time (CfC) cells derived from the Liquid Time-constant ODE, which use a sigmoidal time gate to interpolate between hidden states and embed smooth temporal responses directly into the architecture. Crucially, all other components of the D-3DGS pipeline remain unchanged, making this a near-zero-friction drop-in replacement. Evaluated on eight D-NeRF and seven NeRF-DS benchmark scenes, the liquid field matches or outperforms the MLP baseline in aggregate, with the most notable gains on scenes featuring complex, high-frequency articulated motion. The approach avoids numerical ODE solvers by using the closed-form solution, keeping computational overhead low. The work positions Liquid Neural Networks as a structurally principled alternative for continuous-time modeling in neural rendering pipelines.
What's missing
As a preprint, this work has not yet undergone peer review. The paper does not report computational cost comparisons (training time, inference speed, or parameter count) between the CfC-based and MLP-based deformation fields, nor does it address generalization to multi-camera or non-monocular settings.
What different sources said
- arXiv cs.AICenter
Liquid Neural Networks as a Drop-in Continuous-Time Deformation Field for Dynamic 3D Gaussian Splatting
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.