Deep neural network (DNN) potentials have recently gained popularity in computer simulations of a wide range of molecular systems, from liquids to materials. In this study, we explore the possibility of combining the computational efficiency of the DeePMD framework and the demonstrated accuracy of the MB-pol data-driven, many-body potential to train a DNN potential for large-scale simulations of water across its phase diagram. We find that the DNN potential is able to reliably reproduce the MB-pol results for liquid water, but provides a less accurate description of the vapor–liquid equilibrium properties. This shortcoming is traced back to the inability of the DNN potential to correctly represent many-body interactions. An attempt to explicitly include information about many-body effects results in a new DNN potential that exhibits the opposite performance, being able to correctly reproduce the MB-pol vapor–liquid equilibrium properties, but losing accuracy in the description of the liquid properties. These results suggest that DeePMD-based DNN potentials are not able to correctly “learn” and, consequently, represent many-body interactions, which implies that DNN potentials may have limited ability to predict the properties for state points that are not explicitly included in the training process. The computational efficiency of the DeePMD framework can still be exploited to train DNN potentials on data-driven many-body potentials, which can thus enable large-scale, “chemically accurate” simulations of various molecular systems, with the caveat that the target state points must have been adequately sampled by the reference data-driven many-body potential in order to guarantee a faithful representation of the associated properties.
15 June 2026 in ACM Transactions on Design Automation of Electronic Systems
Recent approaches to congestion prediction and routability-driven placement have made progress, yet important limitations remain. Most existing models rely on heuristic-based intermediate features (e.g., RUDY, pin density) rather than fully end-to-end learning, and many capture either netlist topology or geometric placement context, but not both, limiting their performance. Moreover, when model gradients are used to guide placers, they are often propagated through such intermediate features, reducing their accuracy and effectiveness. To address these gaps, we propose TopoGeoNet , a heterogeneous neural network that unifies graph-based message passing with convolutional U-Net processing over spatial grids. Through bi-directional node–grid communication, TopoGeoNet jointly leverages netlist topology and geometric placement information for accurate congestion prediction. Trained on ISPD’2016 FPGA benchmarks, TopoGeoNet achieves an 8.3% improvement in prediction accuracy over state-of-the-art ML baselines. When integrated into a routability-driven FPGA placer, it improves congestion scores by 16% on average and reduces total place-and-route runtime by 5% without degrading routed wirelength. Compared with state-of-the-art ML prediction models that provide inference and gradients through intermediate features, TopoGeoNet delivers superior routability results, highlighting the advantages of fully end-to-end prediction and differentiability.