The field of dexterous manipulation is moving towards more advanced and realistic robotic hand designs, with a focus on learning from human demonstrations and adapting to new situations. Researchers are exploring the use of novel materials and mechanisms to create more flexible and adaptable robotic hands, such as the use of viscoelastic structures and organic materials. Another area of focus is on developing algorithms and systems that can enable seamless transfer of grasp intent across different robotic hands and morphologies. Noteworthy papers in this area include:

• DexMachina, which proposes a novel curriculum-based algorithm for learning dexterous manipulation policies.
• GEX, which introduces a low-cost dexterous manipulation system that combines a fully-actuated robotic hand with an exoskeleton glove.
• Grasp2Grasp, which presents a vision-based approach to dexterous grasp translation using the Schrödinger Bridge formalism.