CubeSats offer a low-cost platform for space research, particularly for Earth observation. However, their resource-constrained nature and being in space, challenge the flexibility and complexity of the deployed image processing pipelines and their orchestration. This paper introduces two novel systems, DIPP and DISH, to address these challenges. DIPP is a modular and configurable image processing pipeline framework that allows for adaptability to changing mission goals even after deployment, while preserving robustness. DISH is a domain-specific language (DSL) and runtime system designed to schedule complex imaging workloads on low-power and memory-constrained processors. Our experiments demonstrate that DIPP's decomposition of the processing pipelines adds negligible overhead, while significantly reducing the network requirements of updating pipelines and being robust against erroneous module uploads. Furthermore, we compare DISH to Lua, a general purpose scripting language, and demonstrate its comparable expressiveness and lower memory requirement.

While Transformers have revolutionized deep learning, their quadratic attention complexity hinders their ability to process infinitely long inputs. We propose Feedback Attention Memory (FAM), a novel Transformer architecture that leverages a feedback loop to enable the network to attend to its own latent representations. This design fosters the emergence of working memory within the Transformer, allowing it to process indefinitely long sequences. TransformerFAM requires no additional weights, enabling seamless integration with pre-trained models. Our experiments show that TransformerFAM significantly improves Transformer performance on long-context tasks across various model sizes (1B, 8B, and 24B). These results showcase the potential to empower Large Language Models (LLMs) to process sequences of unlimited length.

Memory management spans the layers of computing platforms, ranging across hardware components that map logical addresses to physical memory locations, operating system components that track and control regions (usually in the form of "pages"), and run-time systems and languages that provide simple APIs and/or language constructs to process higher-level objects in terms of lower-level memory segments. An allocation ("new" or "malloc") is performed by placement--choosing a memory address heading a contiguous segment of (at least) a given size. A deallocation ("delete" or "free," perhaps initiated by garbage collectors or other mechanisms that detect unused objects) triggers bookkeeping to enable future memory reuse. Some languages and systems additionally support moving (copying) previously allocated objects to new locations, and automatically adjusting pointers to them accordingly. The primary objective of memory management is fitting all requested memory segments within a given space, or nearly equivalently, minimizing unusable gaps surrounding allocated objects.