Search for: All records

A common failure mode for policies trained with imitation is compounding execution errors at test time. When the learned policy encounters states that are not present in the expert demonstrations, the policy fails, leading to degenerate behavior. The Dataset Aggregation, or DAgger approach to this problem simply collects more data to cover these failure states. However, in practice, this is often prohibitively expensive. In this work, we propose Diffusion Meets DAgger (DMD), a method that reaps the benefits of DAgger but without the cost, for eye-in-hand imitation learning problems. Instead of collecting new samples to cover out-of-distribution states, DMD uses recent advances in diffusion models to synthesize these samples. This leads to robust performance from few demonstrations. We compare DMD against behavior cloning baseline across four tasks: pushing, stacking, pouring, and hanging a shirt. In pushing, DMD achieves 80% success rate with as few as 8 expert demonstrations, where naive behavior cloning reaches only 20%. In stacking, DMD succeeds on average 92% of the time across 5 cups, versus 40% for BC. When pouring coffee beans, DMD transfers to another cup successfully 80% of the time. Finally, DMD attains 90% success rate for hanging shirt on a clothing rack. 

Buildings consume over half of annual energy supply as embodied and operating energy in their construction and operation releasing harmful emissions to the atmosphere. Over 90 % of the embodied energy is attributed to construction materials used in building structure, envelope, and interiors that must be reduced to minimize material use. Concrete is one of the major materials that contributes significantly to the energy and carbon footprint of buildings, as it is responsible for 5-9 % of global carbon emission. Because most of the concrete use in the building sector occurs in building structures, assessing how building design parameters influence its environmental sustainability is important. One of the design parameters that impact the sustainability of buildings is the aspect ratio, which is defined as the ratio of horizontal to vertical surface area of a building. A building with the same floor area can be designed horizontally or vertically with different aspect ratios, which will influence its structural design and eventually the amount of concrete used in the building. In this paper, we examine how aspect ratio may affect the environmental sustainability of a buildings foundation, structural framing, and slab. We model the structure of a generic building with different aspect ratio to analyze if aspect ratio can help reduce the energy and carbon embodied in reinforced concrete structures.