| 1. |
- Krishnan, Sanjay, et al.
(författare)
-
SWIRL : A sequential windowed inverse reinforcement learning algorithm for robot tasks with delayed rewards
- 2019
-
Ingår i: The international journal of robotics research. - : SAGE PUBLICATIONS LTD. - 0278-3649 .- 1741-3176. ; 38:2-3, s. 126-145
-
Tidskriftsartikel (refereegranskat)abstract
- We present sequential windowed inverse reinforcement learning (SWIRL), a policy search algorithm that is a hybrid of exploration and demonstration paradigms for robot learning. We apply unsupervised learning to a small number of initial expert demonstrations to structure future autonomous exploration. SWIRL approximates a long time horizon task as a sequence of local reward functions and subtask transition conditions. Over this approximation, SWIRL applies Q-learning to compute a policy that maximizes rewards. Experiments suggest that SWIRL requires significantly fewer rollouts than pure reinforcement learning and fewer expert demonstrations than behavioral cloning to learn a policy. We evaluate SWIRL in two simulated control tasks, parallel parking and a two-link pendulum. On the parallel parking task, SWIRL achieves the maximum reward on the task with 85% fewer rollouts than Q-learning, and one-eight of demonstrations needed by behavioral cloning. We also consider physical experiments on surgical tensioning and cutting deformable sheets using a da Vinci surgical robot. On the deformable tensioning task, SWIRL achieves a 36% relative improvement in reward compared with a baseline of behavioral cloning with segmentation.
|
|
| 2. |
- Wang, Lixin, et al.
(författare)
-
Doxorubicin-loaded delta inulin conjugates for controlled and targeted drug delivery: Development, characterization, and in vitro evaluation
- 2019
-
Ingår i: Pharmaceutics. - : MDPI AG. - 1999-4923. ; 11:11
-
Tidskriftsartikel (refereegranskat)abstract
- Delta inulin, also known as microparticulate inulin (MPI), was modified by covalently attaching doxorubicin to its nanostructured surface for use as a targeted drug delivery vehicle. MPI is readily endocytosed by monocytes, macrophages, and dendritic cells and in this study, we sought to utilize this property to develop a system to target anti-cancer drugs to lymphoid organs. We investigated, therefore, whether MPI could be used as a vehicle to deliver doxorubicin selectively, thereby reducing the toxicity of this antibiotic anthracycline drug. Doxorubicin was covalently attached to the surface of MPI using an acid-labile linkage to enable pH-controlled release. The MPI-doxorubicin conjugate was characterized using FTIR and SEM, confirming covalent attachment and indicating doxorubicin coupling had no obvious impact on the physical nanostructure, integrity, and cellular uptake of the MPI particles. To simulate the stability of the MPI-doxorubicin in vivo, it was stored in artificial lysosomal fluid (ALF, pH 4.5). Although the MPI-doxorubicin particles were still visible after 165 days in ALF, 53% of glycosidic bonds in the inulin particles were hydrolyzed within 12 days in ALF, reflected by the release of free glucose into solution. By contrast, the fructosidic bonds were much more stable. Drug release studies of the MPI-doxorubicin in vitro, demonstrated a successful pH-dependent controlled release effect. Confocal laser scanning microscopy studies and flow cytometric analysis confirmed that when incubated with live cells, MPI-doxorubicin was effciently internalized by immune cells. An assay of cell metabolic activity demonstrated that the MPI carrier alone had no toxic effects on RAW 264.7 murine monocyte/macrophage-like cells, but exhibited anti-cancer effects against HCT116 human colon cancer cells. MPI-doxorubicin had a greater anti-cancer cell effect than free doxorubicin, particularly when at lower concentrations, suggesting a drug-sparing effect. This study establishes that MPI can be successfully modified with doxorubicin for chemotherapeutic drug delivery.
|
|
