PhD Thesis Defense
PhD Student
Robotics Institute,
Carnegie Mellon University

Redefining the Perception-Action Interface: Visual Action Representations for Contact-Centric Manipulation

GHC 6501

Abstract:  In robotics, understanding the link between perception and action is pivotal. Typically, perception systems process sensory data into state representations like  segmentations and bounding boxes, which a planner uses to plan actions. However, this state estimation approach can fail in environments with partial observability, and in cases with challenging object properties like transparency and deformability.  [...]

Special Events

RI Picnic

Schenley Park Vietnam Veteran's Pavilion , United States

The RI Picnic will be held at the Vietnam Veteran's Pavilion @ Schenley Park on Overlook Drive, Tuesday, August 29, 1-7pm. SOCIALIZE, EAT, DRINK & BE MERRY! Receive this year's RI giveaway item; witness the exciting final rounds of the annual RI croquet tournament; enjoy lawn games right at our own pavilion area. Plan to spend some time at the [...]

RI Event
PhD Student
Robotics Institute,
Carnegie Mellon University

Continual Robot Learning: Benchmarks and Modular Methods

GHC 6501

Abstract: Humans adapt continuously to the world around us, allowing us to acquire new skills and explore diverse environments seamlessly. Current AI methods, however, cannot attain this versatility. Instead, they are typically trained with vast datasets, and learn all tasks simultaneously. However, the trained models have limited ability to adapt to changing contexts, and are [...]

MSR Thesis Defense
PhD Student
Robotics Institute,
Carnegie Mellon University

Architecture and Algorithms for Space-Based Global Wildlife Tracking

GHC 6501

Abstract: Accurate satellite based positioning revolutionized several industries over the past two decades from agriculture to transportation. However, conventional GNSS receivers consume significant amounts of energy and are too large for many applications, including wildlife-tracking which is critical for conservation efforts and improving our understanding of the global climate. To address this capability gap, we [...]

PhD Thesis Defense
PhD Student
Robotics Institute,
Carnegie Mellon University

Multi-Human 3D Reconstruction from Monocular Videos

NSH 4305

Abstract: We study the problem of multi-human 3D reconstruction from videos captured in the wild. Human movements are dynamic, and accurately reconstructing them in various settings is crucial for developing immersive social telepresence, assistive humanoid robots, and augmented reality systems. However, creating such a system requires addressing fundamental issues with previous works regarding the data [...]

MSR Thesis Defense
PhD Student
Robotics Institute,
Carnegie Mellon University

Language-Conditioned Object Detection and Manipulation

NSH 4305

Abstract: Traditional object detection methods are often confined to predefined object vocabularies, limiting their versatility in real-world scenarios where robots need to understand and execute diverse household tasks. Additionally, the 2D and 3D perception communities have typically pursued separate approaches tailored to their respective domains. In this thesis, we present a language-conditioned object detector with [...]

PhD Thesis Defense
PhD Student
Robotics Institute,
Carnegie Mellon University

How I Learned to Love Blobs: The Power of Gaussian Representations in Differentiable Rendering and Optimization

NSH 3305

Abstract: In this thesis, we explore the use of Gaussian Representations in multiple application areas of computer vision and robotics. In particular, we design a ray-based differentiable renderer for 3D Gaussians that can be used to solve multiple classic computer vision problems in a unified manner. For example, we can reconstruct 3D shapes from color, [...]

PhD Thesis Proposal
PhD Student
Robotics Institute,
Carnegie Mellon University

Watch, Practice, Improve: Towards In-the-wild Manipulation

NSH 3305

Abstract: The longstanding dream of many roboticists is to see robots perform diverse tasks in diverse environments. To build such a robot that can operate anywhere, many methods train on robotic interaction data. While these approaches have led to significant advances, they rely on heavily engineered setups or high amounts of supervision, neither of which [...]

VASC Seminar
Aayush Bansal
Startup

Generating Beautiful Pixels

Newell-Simon Hall 3305

Abstract: In this talk, I will present three experiments that use low-level image statistics to generate high-resolution detailed outputs. In the first experiment, I will use 2D pixels to efficiently mine hard examples for better learning. Simply biasing ray sampling towards hard ray examples enables learning of neural fields with more accurate high-frequency detail in less [...]