--- title: 'GRXForm: Molecular Optimization & Form Extraction' url: https://www.emergentmind.com/topics/grxform type: topic --- # GRXForm: Molecular Optimization & Form Extraction GRXForm denotes two distinct systems in recent arXiv literature. In molecular optimization, **GRXForm** abbreviates **“Group-Relative XForm”**, a conditional graph-transformer policy introduced for amortized molecular optimization via Group Relative Policy Optimization (GRPO), with the explicit goal of generating an optimized molecule \(y\) from an input scaffold \(x\) in one forward pass and with **no inference-time oracle calls** [2602.12162]. In a separate information-extraction context, **GRXForm** is also used as the name for a **“Form 57-from-News”** pipeline that populates **Highway-Rail Grade Crossing Incident Data (Form 57)** from news articles in real time [2512.22457]. The primary technical meaning of the term is therefore context-dependent, and the two usages are unrelated except for the shared label. ## 1. Terminological scope and disambiguation Current usage of *GRXForm* is not unique across domains. One usage belongs to molecular design, where the term expands to **Group-Relative XForm** and refers to a scaffold-conditioned policy trained with GRPO. Another usage appears in railway incident information extraction, where the same string names a pipeline for populating **FRA Form 57** from news. A closely related but distinct molecular model is **GraphXForm**, the decoder-only graph transformer that serves as the architectural backbone for the molecular GRXForm system [2602.12162]. | Usage | Domain | Source | |---|---|---| | GRXForm (“Group-Relative XForm”) | Amortized molecular optimization | [2602.12162] | | GRXForm (“Form 57-from-News”) | Railway incident form population from news | [2512.22457] | | GraphXForm | Graph transformer for molecular design | [2411.01667] | The molecular usage is the one for which the name is intrinsic to the method. The railway usage is an explicit project name attached to a pipeline whose paper title is different. This suggests that *GRXForm* is not yet a semantically stabilized identifier across subfields. ## 2. Conditional molecular optimization formulation In the molecular literature, GRXForm is defined by a specific optimization objective: learn a single conditional policy \(\pi_\theta(y \mid x)\) that maps an input scaffold \(x\) to an optimized molecule \(y\), thereby **amortizing the search cost over training** rather than restarting optimization for every input structure [2602.12162]. The motivating contrast is with **“Instance Optimizers”**, which expend significant compute on each new input structure. The workflow is stated compactly as: \[ \text{pre-train on ChEMBL} \rightarrow \text{fine-tune via GRPO on scaffold-constrained tasks} \rightarrow \text{amortized inference: one pass per new scaffold} \rightarrow \text{instant candidate}. \] GRXForm operates over partial molecular graphs. The state at step \(t\) is a partial graph \(G_t=(V_t,E_t)\), and actions are decomposed hierarchically into three levels. **Level 0** chooses the operation in \(\{\text{Stop}, \text{AddAtom}, \text{ModifyExisting}\}\). **Level 1** chooses a target atom \(v\in V_t\). **Level 2** chooses a bond order \(b\in\{1,2,3,4,5,6\}\). At each level, a validity mask \(M(s_t)\) zeroes out choices that would violate valence constraints. The chemical validity guarantee is therefore tied to masked action selection rather than post hoc repair. This conditional formulation matters empirically because the paper identifies a specific failure mode in prior model-based approaches: **high variance arising from the heterogeneous difficulty of distinct starting structures**. GRXForm addresses that variance not by changing the molecular state space, but by changing how rewards are normalized during reinforcement learning. ## 3. Graph-transformer backbone and relation to GraphXForm GRXForm adopts the **decoder-only Graph Transformer from GraphXForm** as its policy backbone [2411.01667]. The predecessor model, GraphXForm, represents each molecule as a **hydrogen-suppressed undirected graph**, uses atom and degree embeddings, omits classical positional encoding to preserve permutation equivariance, and injects bond information through additive attention biases. In the original GraphXForm formulation, node \(0\) is a **virtual atom**; in GRXForm, a **virtual super-node** connects to all atoms for global pooling. For GRXForm, each atom \(i\) is embedded as \[ h_i = \mathrm{AtomTypeEmbed}(\mathrm{type}_i) + \mathrm{DegreeEmbed}(\mathrm{deg}_i) + w_i^{(0)} + w_i^{(1)}, \] where \(w_i^{(\ell)}\) are one-hot indicators recording whether atom \(i\) was selected in sub-action level \(\ell\). The transformer stacks \(L\) self-attention layers with \(H\) heads and no positional encodings. Bond types enter attention through learned scalar biases: \[ A_{ij} \propto \exp\!\left(\frac{Q_i\!\cdot\!K_j}{\sqrt d} + \phi(e_{ij})\right), \] where \(e_{ij}\) is bond order and \(\phi(\cdot)\) is the learned bias. The connection to GraphXForm is architectural and methodological. GraphXForm generates molecules by sequential graph edits, masks invalid actions before renormalization, and reports **100% validity rate by construction** [2411.01667]. It was pretrained on **\(\sim 370\) k ChEMBL molecules** and fine-tuned using a procedure combining **the deep cross-entropy method and self-improvement learning**. GRXForm keeps the graph-editing policy style but replaces that fine-tuning regime with GRPO, shifts the task to scaffold-conditioned amortized optimization, and pre-trains on **\(\sim 1.5\) M ChEMBL35 molecules via teacher-forcing** [2602.12162]. ## 4. Group Relative Policy Optimization The defining algorithmic contribution of GRXForm is **Group Relative Policy Optimization (GRPO)**. Standard REINFORCE with a single global baseline is described as suffering from high variance when scaffold difficulty varies. GRPO addresses this by normalizing rewards **per scaffold group** rather than across the batch [2602.12162]. For a batch of \(B\) scaffolds \(S_1,\dots,S_B\), the policy generates \(G\) completions \(O_{i,1},\dots,O_{i,G}\) for each scaffold \(S_i\), using **Stochastic Beam Search**. If \(r_{i,j}=R(O_{i,j})\), the scaffold-specific mean baseline is \[ \mu_i = \frac{1}{G}\sum_{j=1}^G r_{i,j}, \] and the relative advantage is \[ A_{i,j} = r_{i,j} - \mu_i. \] The policy-gradient estimator is \[ \nabla_\theta J(\theta) \approx \frac{1}{B G} \sum_{i=1}^{B}\sum_{j=1}^{G} A_{i,j} \sum_{t=1}^{T_{i,j}} \nabla_\theta \log \pi_\theta(a_{i,j,t}\mid s_{i,j,