This paper presents a systematic review of a key sector of the much promising and rapidly evolving field of biomedical engineering, specifically on the fabrication of three-dimensional open, porous collagen-based medical devices, using the prominent freeze-drying process. Collagen and its derivatives are the most popular biopolymers in this field, as they constitute the main components of the extracellular matrix, and therefore exhibit desirable properties, such as biocompatibility and biodegradability, for in vivo applications. For this reason, freeze-dried collagen-based sponges with a wide variety of attributes can be produced and have already led to a wide range of successful commercial medical devices, chiefly for dental, orthopedic, hemostatic, and neuronal applications. However, collagen sponges display some vulnerabilities in other key properties, such as low mechanical strength and poor control of their internal architecture, and therefore many studies focus on the settlement of these defects, either by tampering with the steps of the freeze-drying process or by combining collagen with other additives. Furthermore, freeze drying is still considered a high-cost and time-consuming process that is often used in a non-optimized manner. By applying an interdisciplinary approach and combining advances in other technological fields, such as in statistical analysis, implementing the Design of Experiments, and Artificial Intelligence, the opportunity arises to further evolve this process in a sustainable and strategic manner, and optimize the resulting products as well as create new opportunities in this field.
Freeze-drying is a well-established process in biomedical engineering for the fabrication of three-dimensional open-porous medical devices, especially those based on biopolymers. One of the most used biopolymers in this field is collagen, the most abundant protein in the human body and the main component of the extracellular-matrix, as well as its derivatives. Freeze-dried collagen-based sponges with a wide variety of attributes can be produced by design and have led to a wide range of successful commercial medical devices, foremost for dental, orthopedic, hemostatic and neuronal applications. However, this is still considered a high-cost and time-consuming process that is often used in a non-optimized manner. By combining advances in other technological fields, the opportunity arises to further evolve this process in a sustainable manner, and optimize the resulting products as well as create new opportunities in this field.
Graph neural networks remain challenging to apply to traffic forecasting, often failing to consistently outperform simple baselines. We observe that the strongest baseline is regime-dependent: last-value persistence dominates at short horizons on smooth freeway data, while time-of-day climatology prevails at longer horizons and on bursty arterial networks. Rather than treating this as a limitation, we propose a learnable approach that automatically selects the optimal baseline. Our method anchors predictions to a learned, per-horizon convex blend of persistence and climatology, introducing only twelve scalar parameters. This learned anchor recovers whichever baseline is locally most effective, allowing the model to focus on capturing residual variations that neither baseline captures. We evaluate on six public benchmarks (METR-LA, PEMS-BAY, PEMS03/04/07/08) spanning traffic speed and flow data under standard 70/10/20 chronological splits with masked evaluation metrics and holiday-aware climatology. Our anchored temporal models consistently beat both baseline methods on the 12-step average across all datasets, and outperform at every horizon on five of the six benchmarks. When integrated into two strong architectures (STID and Graph WaveNet), the anchor yields substantial gains at long horizons where climatology is most informative. Notably, within our lightweight framework, learned spatial graph components do not improve accuracy and can slightly degrade performance, a finding we analyze and discuss.