想象一下在不久的将来, 在仓库或办公室工作，您 被要求帮助新学员学习其工作的基础知识。捕获:它’是一个机器人。 To teach them, you might want to play a game of “show and tell” — that is, physically showing how to do something a few different ways, while also explaining what you’re doing.
Imagine working at a warehouse or office sometime in the near future, and you’re asked to help a new trainee learn the basics of their job. The catch: It的 a robot. To teach them, you might want to play a game of “show and tell” — that is, physically showing how to do something a few different ways, while also explaining what you’re doing.
根据 Hwang, 的说法，蒙面 IRL 可以帮助机器人在某些环境中安全地进行操作，其中有些元素可能是人类在提示中无法描述的,，但这些元素仍然至关重要。例如,，一台从厨房给你拿零食的机器可能不知道如何避免撞到你的笔记本电脑。同样,，将物品放入不同盒子的工厂机器人必须小心地在货架上导航。
According to Hwang, Masked IRL can help robots safely maneuver in settings where there are elements a human might not describe in a prompt, but that are crucial nonetheless. For example, a machine grabbing you a snack from the kitchen may not know to avoid bumping into your laptop. Likewise, a factory robot placing items into different boxes must carefully navigate around shelves.
为了在这些情况下学习新任务, Masked IRL 使用机器人的 传感器来捕获有关其周围环境的信息。这些组件还记录动觉演示 — 的每个动作，这是一种人类物理移动机器人执行特定动作的训练方法。它的 有点像机器的 物理治疗师, 在特定方向弯曲关节，以向机器人展示如何抓取, 移动, 和放置物体。
To learn new tasks in these situations, Masked IRL uses the robot的 sensors to capture information about its surroundings. These components also log each movement of a kinesthetic demonstration — a training approach where a human physically moves a robot to do a specific action. It的 sort of like being the machine的 physical therapist, bending joints in a particular direction to show a robot how to grab, move, and place objects.
然后，MIT的 系统调用 LLM 将称为轨迹) 的运动序列 ( 与最短可能路径进行比较。该模型还详细说明了提示, 将诸如 “stay close” 之类的请求转换为 “stay close to the surface of the table.” 使用轨迹比较和明确的方向, 中可能不清楚的内容，LLM 开始理解为什么它所训练的动作对任务很重要。然后，第二个法学硕士评估环境, 的细节，例如障碍物的位置和机器人的 目标物体的形状。在此过程中, 它 “masks” (换句话说, 忽略) 它认为与当前任务无关的元素, 将每个元素评分为 “1” (important) 或 “0” (没那么多)。例如，, 无论用户在演示期间是否靠在桌子上，都将是 “0,”，使其无关紧要。任何被视为 “1” 的细节都会通过算法纳入最终的行动计划中。这些面具使 Masked IRL 比 3D 和现实世界演示中的同类基线具有关键优势，因为它教会机器人优先考虑哪些信息。感谢研究人员’系统,，虚拟机器人和真实机器人都能够熟练地绕过障碍物,操纵物体，例如将笔记本电脑周围的咖啡杯移动到桌子上的不同位置。在这些任务中，, Masked IRL 正确识别了用户 偏好,，而他们在提示中没有’t 明确说明,，其频率比可比基线高出 15%。在模拟实验中, CSAIL 研究人员还发现 Masked IRL 学习速度很快。与基线相比，它需要更少的演示来了解如何移动杯子。他们还发现，当 LLM 清除指令, 时，机器人的表现会更好，而不是让机器尝试遵循模糊的请求。这种更有针对性的方法也很好地转化为真正的机械臂,执行系统在训练阶段未见过的’的提示。经过 50 次动觉演示, 训练后，机器人小心地将杯子移向人类，同时避免与用户的计算机 — 发生碰撞。它学会了通过详细说明 “ 远离的更一般要求来避免碰撞这一障碍。” 在 “ 留下时，它还擦拭了桌子靠近” 到它, 并递给用户一袋薯片，同时“远离” 与人和桌子。
MIT的 system then calls on an LLM to compare this sequence of motions (called a trajectory) to the shortest possible path. The model also elaborates on what might be unclear in a prompt, turning a request like “stay close” into “stay close to the surface of the table.” Using the trajectory comparison and clarified directions, the LLM begins to understand why the motions it was trained on are important to the task.
A second LLM then evaluates details of the environment, such as the position of obstacles and the shape of the robot的 target object. During this process, it “masks” (in other words, ignores) the elements it deems irrelevant to the task at hand, scoring each one as either a “1” (important) or “0” (not so much). For example, whether or not a user was leaning on a table during a demonstration would be a “0,” making it irrelevant. Any detail considered a “1” is incorporated into the final action plan by an algorithm.
These masks gave Masked IRL a key advantage over comparable baselines in both 3D and real-world demos because it taught a robot which information to prioritize. Thanks to the researchers system, virtual and real robots alike were able to skillfully maneuver objects around obstacles, such as moving a coffee mug around a laptop to different spots on a table. In these tasks, Masked IRL correctly identified users preferences, which they didn’t explicitly state in their prompts, up to 15 percent more often than comparable baselines.
During simulation experiments, CSAIL researchers also found that Masked IRL was a fast learner. It required fewer demos to understand how to move the mug than its baselines. They also found that the robots performed better when an LLM cleared up instructions, instead of having the machine try to follow a vague request.
This more focused approach also translated well to a real robotic arm, executing prompts the system hadn’t seen during its training phase. After being trained on 50 kinesthetic demonstrations, the robot carefully moved a cup toward a human while avoiding colliding with a user的 computer — an obstacle it learned to avoid by elaborating on a more general request to “stay away.” It also wiped a table down while “staying close” to it, and handed a user a bag of chips while “staying away” from both a human and a table.
蒙面 IRL 感知并解释用户未说的内容,，但很快,它也可能“看到”。 CSAIL 研究人员计划为他们的方法配备摄像头,，让机器人能够拍摄周围环境的图像，从而使他们的方法更加动态。然后它可以突出显示并关注附近的特定元素。例如,，如果您要求机器拾起玩具,，它可能会看到附近有一些香蕉，并在处理目标对象之前忽略它们。
Masked IRL senses and explains what users leave unsaid, but soon, it might “see” it too. CSAIL researchers plan to make their approach more dynamic by equipping it with cameras, allowing a robot to take images of its surroundings. Then it could highlight and focus on specific elements nearby. For example, if you asked the machine to pick up a toy, it might see some bananas nearby and ignore them before handling its target object.