Self-localization is an important technology for automating bulldozers. Conventional bulldozer self-localization systems rely on RTK-GNSS (Real Time Kinematic-Global Navigation Satellite Systems). However, RTK-GNSS signals are sometimes lost in certain mining conditions. Therefore, self-localization methods that do not depend on RTK-GNSS are required. In this paper, we propose a machine learning-based self-localization method for bulldozers. The proposed method consists of two steps: estimating local velocities using a machine learning model from internal sensors, and incorporating these estimates into an Extended Kalman Filter (EKF) for global localization. We also created a novel dataset for bulldozer odometry and conducted experiments across various driving scenarios, including slalom, excavation, and driving on slopes. The result demonstrated that the proposed self-localization method suppressed the accumulation of position errors compared to kinematics-based methods, especially when slip occurred. Furthermore, this study showed that bulldozer-specific sensors, such as blade position sensors and hydraulic pressure sensors, contributed to improving self-localization accuracy.

We address the critical challenge of applying feature attribution methods to the transformer architecture, which dominates current applications in natural language processing and beyond. Traditional attribution methods to explainable AI (XAI) explicitly or implicitly rely on linear or additive surrogate models to quantify the impact of input features on a model's output. In this work, we formally prove an alarming incompatibility: transformers are structurally incapable to align with popular surrogate models for feature attribution, undermining the grounding of these conventional explanation methodologies. To address this discrepancy, we introduce the Softmax-Linked Additive Log-Odds Model (SLALOM), a novel surrogate model specifically designed to align with the transformer framework. Unlike existing methods, SLALOM demonstrates the capacity to deliver a range of faithful and insightful explanations across both synthetic and real-world datasets. Showing that diverse explanations computed from SLALOM outperform common surrogate explanations on different tasks, we highlight the need for task-specific feature attributions rather than a one-size-fits-all approach.

Previously Machine Learning was just seen as an academic exercise within a Computer Science lab or an expensive tool only used within tech heavy companies. But now, as it has become more accessible, it is generating huge excitement from businesses who are looking to use it to solve real practical problems. Here at Slalom we have been privileged to work with a variety of clients and have seen first-hand the benefits that Machine Learning can bring to unlocking the power of their data. This has included identifying new growth opportunities, materially improving customer journeys, identifying efficiencies, and reducing operational risk. However, the most successful implementations we have seen didn't begin with Machine Learning as an aim.

As Machine Learning (ML) gets applied to security-critical or sensitive domains, there is a growing need for integrity and privacy guarantees for ML computations running in untrusted environments. A pragmatic solution comes from Trusted Execution Environments, which use hardware and software protections to isolate sensitive computations from the untrusted software stack. However, these isolation guarantees come at a price in performance, compared to untrusted alternatives. This paper initiates the study of high performance execution of Deep Neural Networks (DNNs) in trusted environments by efficiently partitioning computations between trusted and untrusted devices. Building upon a simple secure outsourcing scheme for matrix multiplication, we propose Slalom, a framework that outsources execution of all linear layers in a DNN from any trusted environment (e.g., SGX, TrustZone or Sanctum) to a faster co-located device. We evaluate Slalom by executing DNNs in an Intel SGX enclave, which selectively outsources work to an untrusted GPU. For two canonical DNNs, VGG16 and MobileNet, we obtain 20x and 6x increases in throughput for verifiable inference, and 10x and 3.5x for verifiable and private inference.

The downhill is Wednesday, so the 22-year-old American suddenly would have had to race on consecutive days. When she tried that earlier at the Pyeongchang Olympics, she followed up her gold in the giant slalom by finishing fourth in the slalom as the defending champion. She pulled out of the super-G on what would have been a third day in a row of racing.

Much has been written about DevOps and its ability to speed up time-to-value and innovation. Machine learning is no different. New approaches and algorithms--for example, deep learning--are coming out all the time, and data scientists are trying them out through code and relying less on GUI-based interfaces. After the new approach has been tested out in a sandbox environment with limited scope, it's time to move toward development, QA, and finally, production. Each one of these environments can be automated with DevOps through tools like Jenkins, Puppet, Chef, Ansible, and Docker.