Large Multimodal Models (LMMs) often rely on in-context learning (ICL) to perform new tasks with minimal supervision. However, ICL performance, especially in smaller LMMs, is inconsistent and does not always improve monotonically with increasing examples. We hypothesize that this occurs due to the LMM being overwhelmed by additional information present in the image embeddings, which is not required for the downstream task. To address this, we propose a meta-learning approach that provides an alternative for inducing few-shot capabilities in LMMs, using a fixed set of soft prompts that are distilled from task-relevant image features and can be adapted at test time using a few examples. To facilitate this distillation, we introduce an attention-mapper module that can be easily integrated with the popular LLaVA v1.5 architecture and is jointly learned with soft prompts, enabling task adaptation in LMMs under low-data regimes with just a few gradient steps. Evaluation on the VL-ICL Bench shows that our method consistently outperforms ICL and related prompt-tuning approaches, even under image perturbations, improving task induction and reasoning across visual question answering tasks.

In environments where many automated guided vehicles (AGVs) operate, planning efficient, collision-free paths is essential. Related research has mainly focused on environments with static passages, resulting in space inefficiency. We define multi-agent and multi-rack path finding (MARPF) as the problem of planning paths for AGVs to convey target racks to their designated locations in environments without passages. In such environments, an AGV without a rack can pass under racks, whereas an AGV with a rack cannot pass under racks to avoid collisions. MARPF entails conveying the target racks without collisions, while the other obstacle racks are positioned without a specific arrangement. AGVs are essential for relocating other racks to prevent any interference with the target racks. We formulated MARPF as an integer linear programming problem in a network flow. To distinguish situations in which an AGV is or is not loading a rack, the proposed method introduces two virtual layers into the network. We optimized the AGVs' movements to move obstacle racks and convey the target racks. The formulation and applicability of the algorithm were validated through numerical experiments. The results indicated that the proposed algorithm addressed issues in environments with dense racks.

Diversity plays a crucial role in improving the performance of multi-agent reinforcement learning (MARL). Currently, many diversity-based methods have been developed to overcome the drawbacks of excessive parameter sharing in traditional MARL. However, there remains a lack of a general metric to quantify policy differences among agents. Such a metric would not only facilitate the evaluation of the diversity evolution in multi-agent systems, but also provide guidance for the design of diversity-based MARL algorithms. In this paper, we propose the multi-agent policy distance (MAPD), a general tool for measuring policy differences in MARL. By learning the conditional representations of agents' decisions, MAPD can computes the policy distance between any pair of agents. Furthermore, we extend MAPD to a customizable version, which can quantify differences among agent policies on specified aspects. Based on the online deployment of MAPD, we design a multi-agent dynamic parameter sharing (MADPS) algorithm as an example of the MAPD's applications. Extensive experiments demonstrate that our method is effective in measuring differences in agent policies and specific behavioral tendencies. Moreover, in comparison to other methods of parameter sharing, MADPS exhibits superior performance.

In automated warehouses, teams of mobile robots fulfill the packaging process by transferring inventory pods to designated workstations while navigating narrow aisles formed by tightly packed pods. This problem is typically modeled as a Multi-Agent Pickup and Delivery (MAPD) problem, which is then solved by repeatedly planning collision-free paths for agents on a fixed graph, as in the Rolling-Horizon Collision Resolution (RHCR) algorithm. However, existing approaches make the limiting assumption that agents are only allowed to move pods that correspond to their current task, while considering the other pods as stationary obstacles (even though all pods are movable). This behavior can result in unnecessarily long paths which could otherwise be avoided by opening additional corridors via pod manipulation. To this end, we explore the implications of allowing agents the flexibility of dynamically relocating pods. We call this new problem Terraforming MAPD (tMAPD) and develop an RHCR-based approach to tackle it. As the extra flexibility of terraforming comes at a significant computational cost, we utilize this capability judiciously by identifying situations where it could make a significant impact on the solution quality. In particular, we invoke terraforming in response to disruptions that often occur in automated warehouses, e.g., when an item is dropped from a pod or when agents malfunction. Empirically, using our approach for tMAPD, where disruptions are modeled via a stochastic process, we improve throughput by over 10%, reduce the maximum service time (the difference between the drop-off time and the pickup time of a pod) by more than 50%, without drastically increasing the runtime, compared to the MAPD setting.

Multi-Agent Pickup and Delivery (MAPD) is the problem of computing collision-free paths for a group of agents such that they can safely reach delivery locations from pickup ones. These locations are provided at runtime, making MAPD a combination between classical Multi-Agent Path Finding (MAPF) and online task assignment. Current algorithms for MAPD do not consider many of the practical issues encountered in real applications: real agents often do not follow the planned paths perfectly, and may be subject to delays and failures. In this paper, we study the problem of MAPD with delays, and we present two solution approaches that provide robustness guarantees by planning paths that limit the effects of imperfect execution. In particular, we introduce two algorithms, k-TP and p-TP, both based on a decentralized algorithm typically used to solve MAPD, Token Passing (TP), which offer deterministic and probabilistic guarantees, respectively. Experimentally, we compare our algorithms against a version of TP enriched with online replanning. k-TP and p-TP provide robust solutions, significantly reducing the number of replans caused by delays, with little or no increase in solution cost and running time.

We propose a distributed planning method with asynchronous execution for multi-agent pickup and delivery (MAPD) problems for environments with occasional delays in agents' activities and flexible endpoints. MAPD is a crucial problem framework with many applications; however, most existing studies assume ideal agent behaviors and environments, such as a fixed speed of agents, synchronized movements, and a well-designed environment with many short detours for multiple agents to perform tasks easily. However, such an environment is often infeasible; for example, the moving speed of agents may be affected by weather and floor conditions and is often prone to delays. The proposed method can relax some infeasible conditions to apply MAPD in more realistic environments by allowing fluctuated speed in agents' actions and flexible working locations (endpoints). Our experiments showed that our method enables agents to perform MAPD in such an environment efficiently, compared to the baseline methods. We also analyzed the behaviors of agents using our method and discuss the limitations.

In this work, we consider the Multi-Agent Pickup-and-Delivery (MAPD) problem, where agents constantly engage with new tasks and need to plan collision-free paths to execute them. To execute a task, an agent needs to visit a pair of goal locations, consisting of a pickup location and a delivery location. We propose two variants of an algorithm that assigns a sequence of tasks to each agent using the anytime algorithm Large Neighborhood Search (LNS) and plans paths using the Multi-Agent Path Finding (MAPF) algorithm Priority-Based Search (PBS). LNS-PBS is complete for well-formed MAPD instances, a realistic subclass of MAPD instances, and empirically more effective than the existing complete MAPD algorithm CENTRAL. LNS-wPBS provides no completeness guarantee but is empirically more efficient and stable than LNS-PBS. It scales to thousands of agents and thousands of tasks in a large warehouse and is empirically more effective than the existing scalable MAPD algorithm HBH+MLA*. LNS-PBS and LNS-wPBS also apply to a more general variant of MAPD, namely the Multi-Goal MAPD (MG-MAPD) problem, where tasks can have different numbers of goal locations.

Dipti Kothari, an intern with MapD, has developed an open source project that uses LiDAR to create a 3D visualization of tree coverage around buildings and the impact of rising sea levels on buildings on Jekyll Island, GA. LiDAR stands for Light Detection and Ranging, a surveying technique which scans a surface using laser light and returns a densely populated 3D point cloud. It lets you answer questions which were previously impractical or impossible to answer. For example, you can calculate the precise tree coverage around every building on an island. Trees are natural windbreakers, lessening the damage caused by hurricanes.

At MapD, we've long been big fans of the PyData stack, and are constantly working on ways for our open source GPU-accelerated analytic SQL engine to play nicely with the terrific tools in the most popular stack that supports open data science. We are founding collaborators of GOAI (the GPU Open Analytics Initiative), working with the awesome folks at Anaconda and H2O.ai, and our friends at NVIDIA. In GOAI, we use Apache Arrow to mediate efficient, high-performance data interchange for analytics and AI workflows. A big reason for doing this is to make MapD itself easily accessible to Python tools. For starters, this means supporting modern Python database interfaces like DBAPI.

A revolution is taking place in the GPU software stack in the fields of analytics, machine learning and deep learning, driven by NVIDIA's hardware innovation, that provides 100x more processing cores and 20x greater memory bandwidth than CPUs. However, systems and platforms are unable to harness these disruptive performance gains because they remain isolated from each other. The GPU Open Analytics Initiative (GOAI) and its first project, the GPU Data Frame (GDF) was created to allow seamless passing of data between processes. At this meetup, we'll explain how we have implemented an end-to-end machine learning powered by GOAI. We will show how GDFs break down the silos to enable interactive data exploration, model training, and model scoring, that is lightning-fast by virtue of avoiding any serialization overhead.