Study Finds Dual-Critic Architecture Outperforms Unified Approach for Humanoid Robot Control
A new study comparing critic architectures for humanoid robot reinforcement learning found that dual critics — using separate reward signals for locomotion and manipulation — substantially outperformed a single unified critic across multiple performance metrics. The research was conducted on a Unitree G1 humanoid robot in simulation, training policies through a 13-level curriculum from stationary reaching to walking with variable targets. The findings suggest critic architecture is a more impactful design choice than reward engineering in multi-objective humanoid RL.
Researchers tested whether humanoid robots learning to simultaneously walk and manipulate objects benefit more from a single unified critic or separate dual critics in a reinforcement learning framework. Using the Unitree G1 humanoid (23 active degrees of freedom) in NVIDIA Isaac Lab, they trained policies through a sequential 13-level curriculum. Dual-critic policies reached targets 3.5 times faster (6.5 vs. 22.6 simulation steps), achieved roughly twice the throughput (14.3 vs. 7.0 validated reaches per 1,000 steps), and attained higher validated reach rates (65.2% vs. 53.8%). Notably, adding anti-gaming reward mechanisms on top of the dual-critic design provided no additional benefit, suggesting the architectural choice itself drives the improvement rather than reward shaping. The authors also highlight implications for RL fine-tuning of imitation-learned policies, warning that a unified critic risks suppressing pre-trained manipulation behaviors through competing locomotion gradients. The paper was accepted at the ICRA 2026 Workshop on Reinforcement Learning for Imitation Learning.
What's missing
All experiments were conducted in simulation (NVIDIA Isaac Lab) with no real-world validation reported; it is unclear whether the dual-critic advantage transfers to physical hardware. The study also does not evaluate generalization beyond the specific 13-level reaching curriculum or to more complex manipulation tasks.
What different sources said
- arXiv cs.LGCenter
Critic Architecture Matters: Dual vs. Unified Critics for Humanoid Loco-Manipulation
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.