This content will become publicly available on May 1, 2024
- Award ID(s):
- NSF-PAR ID:
- Date Published:
- Journal Name:
- International Conference on Robotics and Automation (ICRA)
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
One approach to Imitation Learning is Behavior Cloning, in which a robot observes a supervisor and infers a control policy. A known problem with this “off-policy” approach is that the robot’s errors compound when drifting away from the supervisor’s demonstrations. On-policy, techniques alleviate this by iteratively collecting corrective actions for the current robot policy. However, these techniques can be tedious for human supervisors, add significant computation burden, and may visit dangerous states during training. We propose an off-policy approach that injects noise into the supervisor’s policy while demonstrating. This forces the supervisor to demonstrate how to recover from errors. We propose a new algorithm, DART (Disturbances for Augmenting Robot Trajectories), that collects demonstrations with injected noise, and optimizes the noise level to approximate the error of the robot’s trained policy during data collection. We compare DART with DAgger and Behavior Cloning in two domains: in simulation with an algorithmic supervisor on the MuJoCo tasks (Walker, Humanoid, Hopper, Half-Cheetah) and in physical experiments with human supervisors training a Toyota HSR robot to perform grasping in clutter. For high dimensional tasks like Humanoid, DART can be up to 3x faster in computation time and only decreases the supervisor’s cumulative reward by 5% during training, whereas DAgger executes policies that have 80% less cumulative reward than the supervisor. On the grasping in clutter task, DART obtains on average a 62% performance increase over Behavior Cloning.more » « less
Obeid, I. ; Selesnik, I. ; Picone, J. (Ed.)The Neuronix high-performance computing cluster allows us to conduct extensive machine learning experiments on big data . This heterogeneous cluster uses innovative scheduling technology, Slurm , that manages a network of CPUs and graphics processing units (GPUs). The GPU farm consists of a variety of processors ranging from low-end consumer grade devices such as the Nvidia GTX 970 to higher-end devices such as the GeForce RTX 2080. These GPUs are essential to our research since they allow extremely compute-intensive deep learning tasks to be executed on massive data resources such as the TUH EEG Corpus . We use TensorFlow  as the core machine learning library for our deep learning systems, and routinely employ multiple GPUs to accelerate the training process. Reproducible results are essential to machine learning research. Reproducibility in this context means the ability to replicate an existing experiment – performance metrics such as error rates should be identical and floating-point calculations should match closely. Three examples of ways we typically expect an experiment to be replicable are: (1) The same job run on the same processor should produce the same results each time it is run. (2) A job run on a CPU and GPU should produce identical results. (3) A job should produce comparable results if the data is presented in a different order. System optimization requires an ability to directly compare error rates for algorithms evaluated under comparable operating conditions. However, it is a difficult task to exactly reproduce the results for large, complex deep learning systems that often require more than a trillion calculations per experiment . This is a fairly well-known issue and one we will explore in this poster. Researchers must be able to replicate results on a specific data set to establish the integrity of an implementation. They can then use that implementation as a baseline for comparison purposes. A lack of reproducibility makes it very difficult to debug algorithms and validate changes to the system. Equally important, since many results in deep learning research are dependent on the order in which the system is exposed to the data, the specific processors used, and even the order in which those processors are accessed, it becomes a challenging problem to compare two algorithms since each system must be individually optimized for a specific data set or processor. This is extremely time-consuming for algorithm research in which a single run often taxes a computing environment to its limits. Well-known techniques such as cross-validation [5,6] can be used to mitigate these effects, but this is also computationally expensive. These issues are further compounded by the fact that most deep learning algorithms are susceptible to the way computational noise propagates through the system. GPUs are particularly notorious for this because, in a clustered environment, it becomes more difficult to control which processors are used at various points in time. Another equally frustrating issue is that upgrades to the deep learning package, such as the transition from TensorFlow v1.9 to v1.13, can also result in large fluctuations in error rates when re-running the same experiment. Since TensorFlow is constantly updating functions to support GPU use, maintaining an historical archive of experimental results that can be used to calibrate algorithm research is quite a challenge. This makes it very difficult to optimize the system or select the best configurations. The overall impact of all of these issues described above is significant as error rates can fluctuate by as much as 25% due to these types of computational issues. Cross-validation is one technique used to mitigate this, but that is expensive since you need to do multiple runs over the data, which further taxes a computing infrastructure already running at max capacity. GPUs are preferred when training a large network since these systems train at least two orders of magnitude faster than CPUs . Large-scale experiments are simply not feasible without using GPUs. However, there is a tradeoff to gain this performance. Since all our GPUs use the NVIDIA CUDA® Deep Neural Network library (cuDNN) , a GPU-accelerated library of primitives for deep neural networks, it adds an element of randomness into the experiment. When a GPU is used to train a network in TensorFlow, it automatically searches for a cuDNN implementation. NVIDIA’s cuDNN implementation provides algorithms that increase the performance and help the model train quicker, but they are non-deterministic algorithms [9,10]. Since our networks have many complex layers, there is no easy way to avoid this randomness. Instead of comparing each epoch, we compare the average performance of the experiment because it gives us a hint of how our model is performing per experiment, and if the changes we make are efficient. In this poster, we will discuss a variety of issues related to reproducibility and introduce ways we mitigate these effects. For example, TensorFlow uses a random number generator (RNG) which is not seeded by default. TensorFlow determines the initialization point and how certain functions execute using the RNG. The solution for this is seeding all the necessary components before training the model. This forces TensorFlow to use the same initialization point and sets how certain layers work (e.g., dropout layers). However, seeding all the RNGs will not guarantee a controlled experiment. Other variables can affect the outcome of the experiment such as training using GPUs, allowing multi-threading on CPUs, using certain layers, etc. To mitigate our problems with reproducibility, we first make sure that the data is processed in the same order during training. Therefore, we save the data from the last experiment and to make sure the newer experiment follows the same order. If we allow the data to be shuffled, it can affect the performance due to how the model was exposed to the data. We also specify the float data type to be 32-bit since Python defaults to 64-bit. We try to avoid using 64-bit precision because the numbers produced by a GPU can vary significantly depending on the GPU architecture [11-13]. Controlling precision somewhat reduces differences due to computational noise even though technically it increases the amount of computational noise. We are currently developing more advanced techniques for preserving the efficiency of our training process while also maintaining the ability to reproduce models. In our poster presentation we will demonstrate these issues using some novel visualization tools, present several examples of the extent to which these issues influence research results on electroencephalography (EEG) and digital pathology experiments and introduce new ways to manage such computational issues.more » « less
In the realm of robotics and automation, robot teleoperation, which facilitates human–machine interaction in distant or hazardous settings, has surged in significance. A persistent issue in this domain is the delays between command issuance and action execution, causing negative repercussions on operator situational awareness, performance, and cognitive load. These delays, particularly in long-distance operations, are difficult to mitigate even with the most advanced computing advancements. Current solutions mainly revolve around machine-based adjustments to combat these delays. However, a notable lacuna remains in harnessing human perceptions for an enhanced subjective teleoperation experience. This paper introduces a novel approach of sensory manipulation for induced human adaptation in delayed teleoperation. Drawing from motor learning and rehabilitation principles, it is posited that strategic sensory manipulation, via altered sensory stimuli, can mitigate the subjective feeling of these delays. The focus is not on introducing new skills or adapting to novel conditions; rather, it leverages prior motor coordination experience in the context of delays. The objective is to reduce the need for extensive training or sophisticated automation designs. A human-centered experiment involving 41 participants was conducted to examine the effects of modified haptic cues in teleoperations with delays. These cues were generated from high-fidelity physics engines using parameters from robot-end sensors or physics engine simulations. The results underscored several benefits, notably the considerable reduction in task time and enhanced user perceptions about visual delays. Real-time haptic feedback, or the anchoring method, emerged as a significant contributor to these benefits, showcasing reduced cognitive load, bolstered self-confidence, and minimized frustration. Beyond the prevalent methods of automation design and training, this research underscores induced human adaptation as a pivotal avenue in robot teleoperation. It seeks to enhance teleoperation efficacy through rapid human adaptation, offering insights beyond just optimizing robotic systems for delay compensations.
Successful surgical operations are characterized by preplanning routines to be executed during actual surgical operations. To achieve this, surgeons rely on the experience acquired from the use of cadavers, enabling technologies like virtual reality (VR) and clinical years of practice. However, cadavers, having no dynamism and realism as they lack blood, can exhibit limited tissue degradation and shrinkage, while current VR systems do not provide amplified haptic feedback. This can impact surgical training increasing the likelihood of medical errors. This work proposes a novel Mixed Reality Combination System (MRCS) that pairs Augmented Reality (AR) technology and an inertial measurement unit (IMU) sensor with 3D printed, collagen-based specimens that can enhance task performance like planning and execution. To achieve this, the MRCS charts out a path prior to a user task execution based on a visual, physical, and dynamic environment on the state of a target object by utilizing surgeon-created virtual imagery that, when projected onto a 3D printed biospecimen as AR, reacts visually to user input on its actual physical state. This allows a real-time user reaction of the MRCS by displaying new multi-sensory virtual states of an object prior to performing on the actual physical state of that same object enabling effective task planning. Tracked user actions using an integrated 9-Degree of Freedom IMU demonstrate task execution This demonstrates that a user, with limited knowledge of specific anatomy, can, under guidance, execute a preplanned task. In addition, to surgical planning, this system can be generally applied in areas such as construction, maintenance, and education.
Current textual question answering (QA) models achieve strong performance on in-domain test sets, but often do so by fitting surface-level patterns, so they fail to generalize to out-of-distribution settings. To make a more robust and understandable QA system, we model question answering as an alignment problem. We decompose both the question and context into smaller units based on off-the-shelf semantic representations (here, semantic roles), and align the question to a subgraph of the context in order to find the answer. We formulate our model as a structured SVM, with alignment scores computed via BERT, and we can train end-to-end despite using beam search for approximate inference. Our use of explicit alignments allows us to explore a set of constraints with which we can prohibit certain types of bad model behavior arising in cross-domain settings. Furthermore, by investigating differences in scores across different potential answers, we can seek to understand what particular aspects of the input lead the model to choose the answer without relying on post-hoc explanation techniques. We train our model on SQuAD v1.1 and test it on several adversarial and out-of-domain datasets. The results show that our model is more robust than the standard BERT QA model, and constraints derived from alignment scores allow us to effectively trade off coverage and accuracy.more » « less