Search

With the widespread deployment of Multimodal Large Language Models (MLLMs) for visual-reasoning tasks, improving their safety has become crucial. Recent research indicates that despite training-time safety alignment, these models remain vulnerable to jailbreak attacks: carefully crafted image-prompt pairs that compel the model to generate harmful content. In this work, we first highlight a critical safety gap, demonstrating that alignment achieved solely through safety training may be insufficient against jailbreak attacks. To address this vulnerability, we propose Immune, an inference-time defense framework that leverages a safe reward model during decoding to defend against jailbreak attacks. Additionally, we provide a rigorous mathematical characterization of Immune, offering provable guarantees against jailbreaks. Extensive evaluations on diverse jailbreak benchmarks using recent MLLMs reveal that Immune effectively enhances model safety while preserving the model's original capabilities. For instance, against text-based jailbreak attacks on LLaVA-1.6, Immune reduces the attack success rate by 57.82% and 16.78% compared to the base MLLM and state-of-the-art defense strategy, respectively.

The ability to comprehend audio--which includes speech, non-speech sounds, and music--is crucial for AI agents to interact effectively with the world. We present MMAU, a novel benchmark designed to evaluate multimodal audio understanding models on tasks requiring expert-level knowledge and complex reasoning. MMAU comprises 10k carefully curated audio clips paired with human-annotated natural language questions and answers spanning speech, environmental sounds, and music. It includes information extraction and reasoning questions, requiring models to demonstrate 27 distinct skills across unique and challenging tasks. Unlike existing benchmarks, MMAU emphasizes advanced perception and reasoning with domain-specific knowledge, challenging models to tackle tasks akin to those faced by experts. We assess 18 open-source and proprietary (Large) Audio-Language Models, demonstrating the significant challenges posed by MMAU. Notably, even the most advanced Gemini Pro v1.5 achieves only 52.97% accuracy, and the state-of-the-art open-source Qwen2-Audio achieves only 52.50%, highlighting considerable room for improvement. We believe MMAU will drive the audio and multimodal research community to develop more advanced audio understanding models capable of solving complex audio tasks., Comment: Project Website: https://sakshi113.github.io/mmau_homepage/

Open-vocabulary audio language models (ALMs), like Contrastive Language Audio Pretraining (CLAP), represent a promising new paradigm for audio-text retrieval using natural language queries. In this paper, for the first time, we perform controlled experiments on various benchmarks to show that existing ALMs struggle to generalize to linguistic variations in textual queries. To address this issue, we propose RobustCLAP, a novel and compute-efficient technique to learn audio-language representations agnostic to linguistic variations. Specifically, we reformulate the contrastive loss used in CLAP architectures by introducing a multi-view contrastive learning objective, where paraphrases are treated as different views of the same audio scene and use this for training. Our proposed approach improves the text-to-audio retrieval performance of CLAP by 0.8%-13% across benchmarks and enhances robustness to linguistic variation., Comment: 15 pages

Document structure editing involves manipulating localized textual, visual, and layout components in document images based on the user's requests. Past works have shown that multimodal grounding of user requests in the document image and identifying the accurate structural components and their associated attributes remain key challenges for this task. To address these, we introduce the DocEdit-v2, a novel framework that performs end-to-end document editing by leveraging Large Multimodal Models (LMMs). It consists of three novel components: (1) Doc2Command, which simultaneously localizes edit regions of interest (RoI) and disambiguates user edit requests into edit commands; (2) LLM-based Command Reformulation prompting to tailor edit commands originally intended for specialized software into edit instructions suitable for generalist LMMs. (3) Moreover, DocEdit-v2 processes these outputs via Large Multimodal Models like GPT-4V and Gemini, to parse the document layout, execute edits on grounded Region of Interest (RoI), and generate the edited document image. Extensive experiments on the DocEdit dataset show that DocEdit-v2 significantly outperforms strong baselines on edit command generation (2-33%), RoI bounding box detection (12-31%), and overall document editing (1-12\%) tasks., Comment: EMNLP 2024 (Main)

Audio-Language Models (ALMs) have demonstrated remarkable performance in zero-shot audio classification. In this paper, we introduce PAT (Parameter-free Audio-Text aligner), a simple and training-free method aimed at boosting the zero-shot audio classification performance of CLAP-like ALMs. To achieve this, we propose to improve the cross-modal interaction between audio and language modalities by enhancing the representations for both modalities using mutual feedback. Precisely, to enhance textual representations, we propose a prompt ensemble algorithm that automatically selects and combines the most relevant prompts from a datastore with a large pool of handcrafted prompts and weighs them according to their relevance to the audio. On the other hand, to enhance audio representations, we reweigh the frame-level audio features based on the enhanced textual information. Our proposed method does not require any additional modules or parameters and can be used with any existing CLAP-like ALM to improve zero-shot audio classification performance. We experiment across 18 diverse benchmark datasets and 6 ALMs and show that the PAT outperforms vanilla zero-shot evaluation with significant margins of 0.42%-27.0%. Additionally, we demonstrate that PAT maintains robust performance even when input audio is degraded by varying levels of noise. Our code will be open-sourced upon acceptance., Comment: 18 pages

Generative Error Correction (GEC) has emerged as a powerful post-processing method to enhance the performance of Automatic Speech Recognition (ASR) systems. However, we show that GEC models struggle to generalize beyond the specific types of errors encountered during training, limiting their ability to correct new, unseen errors at test time, particularly in out-of-domain (OOD) scenarios. This phenomenon amplifies with named entities (NEs), where, in addition to insufficient contextual information or knowledge about the NEs, novel NEs keep emerging. To address these issues, we propose DARAG (Data- and Retrieval-Augmented Generative Error Correction), a novel approach designed to improve GEC for ASR in in-domain (ID) and OOD scenarios. We augment the GEC training dataset with synthetic data generated by prompting LLMs and text-to-speech models, thereby simulating additional errors from which the model can learn. For OOD scenarios, we simulate test-time errors from new domains similarly and in an unsupervised fashion. Additionally, to better handle named entities, we introduce retrieval-augmented correction by augmenting the input with entities retrieved from a database. Our approach is simple, scalable, and both domain- and language-agnostic. We experiment on multiple datasets and settings, showing that DARAG outperforms all our baselines, achieving 8\% -- 30\% relative WER improvements in ID and 10\% -- 33\% improvements in OOD settings., Comment: Preprint. Under Review

In this paper, we present EH-MAM (Easy-to-Hard adaptive Masked Acoustic Modeling), a novel self-supervised learning approach for speech representation learning. In contrast to the prior methods that use random masking schemes for Masked Acoustic Modeling (MAM), we introduce a novel selective and adaptive masking strategy. Specifically, during SSL training, we progressively introduce harder regions to the model for reconstruction. Our approach automatically selects hard regions and is built on the observation that the reconstruction loss of individual frames in MAM can provide natural signals to judge the difficulty of solving the MAM pre-text task for that frame. To identify these hard regions, we employ a teacher model that first predicts the frame-wise losses and then decides which frames to mask. By learning to create challenging problems, such as identifying harder frames and solving them simultaneously, the model is able to learn more effective representations and thereby acquire a more comprehensive understanding of the speech. Quantitatively, EH-MAM outperforms several state-of-the-art baselines across various low-resource speech recognition and SUPERB benchmarks by 5%-10%. Additionally, we conduct a thorough analysis to show that the regions masked by EH-MAM effectively capture useful context across speech frames.

We introduce ET-Former, a novel end-to-end algorithm for semantic scene completion using a single monocular camera. Our approach generates a semantic occupancy map from single RGB observation while simultaneously providing uncertainty estimates for semantic predictions. By designing a triplane-based deformable attention mechanism, our approach improves geometric understanding of the scene than other SOTA approaches and reduces noise in semantic predictions. Additionally, through the use of a Conditional Variational AutoEncoder (CVAE), we estimate the uncertainties of these predictions. The generated semantic and uncertainty maps will aid in the formulation of navigation strategies that facilitate safe and permissible decision-making in the future. Evaluated on the Semantic-KITTI dataset, ET-Former achieves the highest IoU and mIoU, surpassing other methods by 15.16% in IoU and 24.24% in mIoU, while reducing GPU memory usage of existing methods by 25%-50.5%.

Recently, 3D Gaussian splatting has gained attention for its capability to generate high-fidelity rendering results. At the same time, most applications such as games, animation, and AR/VR use mesh-based representations to represent and render 3D scenes. We propose a novel approach that integrates mesh representation with 3D Gaussian splats to perform high-quality rendering of reconstructed real-world scenes. In particular, we introduce a distance-based Gaussian splatting technique to align the Gaussian splats with the mesh surface and remove redundant Gaussian splats that do not contribute to the rendering. We consider the distance between each Gaussian splat and the mesh surface to distinguish between tightly-bound and loosely-bound Gaussian splats. The tightly-bound splats are flattened and aligned well with the mesh geometry. The loosely-bound Gaussian splats are used to account for the artifacts in reconstructed 3D meshes in terms of rendering. We present a training strategy of binding Gaussian splats to the mesh geometry, and take into account both types of splats. In this context, we introduce several regularization techniques aimed at precisely aligning tightly-bound Gaussian splats with the mesh surface during the training process. We validate the effectiveness of our method on large and unbounded scene from mip-NeRF 360 and Deep Blending datasets. Our method surpasses recent mesh-based neural rendering techniques by achieving a 2dB higher PSNR, and outperforms mesh-based Gaussian splatting methods by 1.3 dB PSNR, particularly on the outdoor mip-NeRF 360 dataset, demonstrating better rendering quality. We provide analyses for each type of Gaussian splat and achieve a reduction in the number of Gaussian splats by 30% compared to the original 3D Gaussian splatting., Comment: ACCV (Asian Conference on Computer Vision) 2024

We present a novel-view rendering algorithm, Mode-GS, for ground-robot trajectory datasets. Our approach is based on using anchored Gaussian splats, which are designed to overcome the limitations of existing 3D Gaussian splatting algorithms. Prior neural rendering methods suffer from severe splat drift due to scene complexity and insufficient multi-view observation, and can fail to fix splats on the true geometry in ground-robot datasets. Our method integrates pixel-aligned anchors from monocular depths and generates Gaussian splats around these anchors using residual-form Gaussian decoders. To address the inherent scale ambiguity of monocular depth, we parameterize anchors with per-view depth-scales and employ scale-consistent depth loss for online scale calibration. Our method results in improved rendering performance, based on PSNR, SSIM, and LPIPS metrics, in ground scenes with free trajectory patterns, and achieves state-of-the-art rendering performance on the R3LIVE odometry dataset and the Tanks and Temples dataset.

Text-based AI system optimization typically involves a feedback loop scheme where a single LLM generates an evaluation in natural language of the current output to improve the next iteration's output. However, in this work, we empirically demonstrate that for a practical and complex task (code generation) with multiple criteria to evaluate, utilizing only one LLM evaluator tends to let errors in generated code go undetected, thus leading to incorrect evaluations and ultimately suboptimal test case performance. Motivated by this failure case, we assume there exists an optimal evaluation policy that samples an evaluation between response and ground truth. We then theoretically prove that a linear combination of multiple evaluators can approximate this optimal policy. From this insight, we propose AI system optimization via Multiple LLM Evaluators (AIME). AIME is an evaluation protocol that utilizes multiple LLMs that each independently generate an evaluation on separate criteria and then combine them via concatenation. We provide an extensive empirical study showing AIME outperforming baseline methods in code generation tasks, with up to $62\%$ higher error detection rate and up to $16\%$ higher success rate than a single LLM evaluation protocol on LeetCodeHard and HumanEval datasets. We also show that the selection of the number of evaluators and which criteria to utilize is non-trivial as it can impact pact success rate by up to $12\%$., Comment: 21 pages, 10 Figures, 4 Tables

We present Synthio, a novel approach for augmenting small-scale audio classification datasets with synthetic data. Our goal is to improve audio classification accuracy with limited labeled data. Traditional data augmentation techniques, which apply artificial transformations (e.g., adding random noise or masking segments), struggle to create data that captures the true diversity present in real-world audios. To address this shortcoming, we propose to augment the dataset with synthetic audio generated from text-to-audio (T2A) diffusion models. However, synthesizing effective augmentations is challenging because not only should the generated data be acoustically consistent with the underlying small-scale dataset, but they should also have sufficient compositional diversity. To overcome the first challenge, we align the generations of the T2A model with the small-scale dataset using preference optimization. This ensures that the acoustic characteristics of the generated data remain consistent with the small-scale dataset. To address the second challenge, we propose a novel caption generation technique that leverages the reasoning capabilities of Large Language Models to (1) generate diverse and meaningful audio captions and (2) iteratively refine their quality. The generated captions are then used to prompt the aligned T2A model. We extensively evaluate Synthio on ten datasets and four simulated limited-data settings. Results indicate our method consistently outperforms all baselines by 0.1%-39% using a T2A model trained only on weakly-captioned AudioSet., Comment: Code and Checkpoints will be soon available here: https://github.com/Sreyan88/Synthio

We present a novel autonomous robot navigation algorithm for outdoor environments that is capable of handling diverse terrain traversability conditions. Our approach, VLM-GroNav, uses vision-language models (VLMs) and integrates them with physical grounding that is used to assess intrinsic terrain properties such as deformability and slipperiness. We use proprioceptive-based sensing, which provides direct measurements of these physical properties, and enhances the overall semantic understanding of the terrains. Our formulation uses in-context learning to ground the VLM's semantic understanding with proprioceptive data to allow dynamic updates of traversability estimates based on the robot's real-time physical interactions with the environment. We use the updated traversability estimations to inform both the local and global planners for real-time trajectory replanning. We validate our method on a legged robot (Ghost Vision 60) and a wheeled robot (Clearpath Husky), in diverse real-world outdoor environments with different deformable and slippery terrains. In practice, we observe significant improvements over state-of-the-art methods by up to 50% increase in navigation success rate.

We introduce SOAR, a novel Self-supervised pretraining algorithm for aerial footage captured by Unmanned Aerial Vehicles (UAVs). We incorporate human object knowledge throughout the pretraining process to enhance UAV video pretraining efficiency and downstream action recognition performance. This is in contrast to prior works that primarily incorporate object information during the fine-tuning stage. Specifically, we first propose a novel object-aware masking strategy designed to retain the visibility of certain patches related to objects throughout the pretraining phase. Second, we introduce an object-aware loss function that utilizes object information to adjust the reconstruction loss, preventing bias towards less informative background patches. In practice, SOAR with a vanilla ViT backbone, outperforms best UAV action recognition models, recording a 9.7% and 21.4% boost in top-1 accuracy on the NEC-Drone and UAV-Human datasets, while delivering an inference speed of 18.7ms per video, making it 2x to 5x faster. Additionally, SOAR obtains comparable accuracy to prior self-supervised learning (SSL) methods while requiring 87.5% less pretraining time and 25% less memory usage

We present CROSS-GAiT, a novel algorithm for quadruped robots that uses Cross Attention to fuse terrain representations derived from visual and time-series inputs, including linear accelerations, angular velocities, and joint efforts. These fused representations are used to adjust the robot's step height and hip splay, enabling adaptive gaits that respond dynamically to varying terrain conditions. We generate these terrain representations by processing visual inputs through a masked Vision Transformer (ViT) encoder and time-series data through a dilated causal convolutional encoder. The cross-attention mechanism then selects and integrates the most relevant features from each modality, combining terrain characteristics with robot dynamics for better-informed gait adjustments. CROSS-GAiT uses the combined representation to dynamically adjust gait parameters in response to varying and unpredictable terrains. We train CROSS-GAiT on data from diverse terrains, including asphalt, concrete, brick pavements, grass, dense vegetation, pebbles, gravel, and sand. Our algorithm generalizes well and adapts to unseen environmental conditions, enhancing real-time navigation performance. CROSS-GAiT was implemented on a Ghost Robotics Vision 60 robot and extensively tested in complex terrains with high vegetation density, uneven/unstable surfaces, sand banks, deformable substrates, etc. We observe at least a 7.04% reduction in IMU energy density and a 27.3% reduction in total joint effort, which directly correlates with increased stability and reduced energy usage when compared to state-of-the-art methods. Furthermore, CROSS-GAiT demonstrates at least a 64.5% increase in success rate and a 4.91% reduction in time to reach the goal in four complex scenarios. Additionally, the learned representations perform 4.48% better than the state-of-the-art on a terrain classification task.

We present BehAV, a novel approach for autonomous robot navigation in outdoor scenes guided by human instructions and leveraging Vision Language Models (VLMs). Our method interprets human commands using a Large Language Model (LLM) and categorizes the instructions into navigation and behavioral guidelines. Navigation guidelines consist of directional commands (e.g., "move forward until") and associated landmarks (e.g., "the building with blue windows"), while behavioral guidelines encompass regulatory actions (e.g., "stay on") and their corresponding objects (e.g., "pavements"). We use VLMs for their zero-shot scene understanding capabilities to estimate landmark locations from RGB images for robot navigation. Further, we introduce a novel scene representation that utilizes VLMs to ground behavioral rules into a behavioral cost map. This cost map encodes the presence of behavioral objects within the scene and assigns costs based on their regulatory actions. The behavioral cost map is integrated with a LiDAR-based occupancy map for navigation. To navigate outdoor scenes while adhering to the instructed behaviors, we present an unconstrained Model Predictive Control (MPC)-based planner that prioritizes both reaching landmarks and following behavioral guidelines. We evaluate the performance of BehAV on a quadruped robot across diverse real-world scenarios, demonstrating a 22.49% improvement in alignment with human-teleoperated actions, as measured by Frechet distance, and achieving a 40% higher navigation success rate compared to state-of-the-art methods.

We present ImPoster, a novel algorithm for generating a target image of a 'source' subject performing a 'driving' action. The inputs to our algorithm are a single pair of a source image with the subject that we wish to edit and a driving image with a subject of an arbitrary class performing the driving action, along with the text descriptions of the two images. Our approach is completely unsupervised and does not require any access to additional annotations like keypoints or pose. Our approach builds on a pretrained text-to-image latent diffusion model and learns the characteristics of the source and the driving image by finetuning the diffusion model for a small number of iterations. At inference time, ImPoster performs step-wise text prompting i.e. it denoises by first moving in the direction of the image manifold corresponding to the driving image followed by the direction of the image manifold corresponding to the text description of the desired target image. We propose a novel diffusion guidance formulation, image frequency guidance, to steer the generation towards the manifold of the source subject and the driving action at every step of the inference denoising. Our frequency guidance formulations are derived from the frequency domain properties of images. We extensively evaluate ImPoster on a diverse set of source-driving image pairs to demonstrate improvements over baselines. To the best of our knowledge, ImPoster is the first approach towards achieving both subject-driven as well as action-driven image personalization. Code and data is available at https://github.com/divyakraman/ImPosterDiffusion2024.

Navigating large-scale outdoor environments requires complex reasoning in terms of geometric structures, environmental semantics, and terrain characteristics, which are typically captured by onboard sensors such as LiDAR and cameras. While current mobile robots can navigate such environments using pre-defined, high-precision maps based on hand-crafted rules catered for the specific environment, they lack commonsense reasoning capabilities that most humans possess when navigating unknown outdoor spaces. To address this gap, we introduce the Global Navigation Dataset (GND), a large-scale dataset that integrates multi-modal sensory data, including 3D LiDAR point clouds and RGB and 360-degree images, as well as multi-category traversability maps (pedestrian walkways, vehicle roadways, stairs, off-road terrain, and obstacles) from ten university campuses. These environments encompass a variety of parks, urban settings, elevation changes, and campus layouts of different scales. The dataset covers approximately 2.7km2 and includes at least 350 buildings in total. We also present a set of novel applications of GND to showcase its utility to enable global robot navigation, such as map-based global navigation, mapless navigation, and global place recognition.

Open-vocabulary audio-language models, like CLAP, offer a promising approach for zero-shot audio classification (ZSAC) by enabling classification with any arbitrary set of categories specified with natural language prompts. In this paper, we propose a simple but effective method to improve ZSAC with CLAP. Specifically, we shift from the conventional method of using prompts with abstract category labels (e.g., Sound of an organ) to prompts that describe sounds using their inherent descriptive features in a diverse context (e.g.,The organ's deep and resonant tones filled the cathedral.). To achieve this, we first propose ReCLAP, a CLAP model trained with rewritten audio captions for improved understanding of sounds in the wild. These rewritten captions describe each sound event in the original caption using their unique discriminative characteristics. ReCLAP outperforms all baselines on both multi-modal audio-text retrieval and ZSAC. Next, to improve zero-shot audio classification with ReCLAP, we propose prompt augmentation. In contrast to the traditional method of employing hand-written template prompts, we generate custom prompts for each unique label in the dataset. These custom prompts first describe the sound event in the label and then employ them in diverse scenes. Our proposed method improves ReCLAP's performance on ZSAC by 1%-18% and outperforms all baselines by 1% - 55%., Comment: Code and Checkpoints: https://github.com/Sreyan88/ReCLAP

We present VLPG-Nav, a visual language navigation method for guiding robots to specified objects within household scenes. Unlike existing methods primarily focused on navigating the robot toward objects, our approach considers the additional challenge of centering the object within the robot's camera view. Our method builds a visual language pose graph (VLPG) that functions as a spatial map of VL embeddings. Given an open vocabulary object query, we plan a viewpoint for object navigation using the VLPG. Despite navigating to the viewpoint, real-world challenges like object occlusion, displacement, and the robot's localization error can prevent visibility. We build an object localization probability map that leverages the robot's current observations and prior VLPG. When the object isn't visible, the probability map is updated and an alternate viewpoint is computed. In addition, we propose an object-centering formulation that locally adjusts the robot's pose to center the object in the camera view. We evaluate the effectiveness of our approach through simulations and real-world experiments, evaluating its ability to successfully view and center the object within the camera field of view. VLPG-Nav demonstrates improved performance in locating the object, navigating around occlusions, and centering the object within the robot's camera view, outperforming the selected baselines in the evaluation metrics.

Recent 3D novel view synthesis (NVS) methods often require extensive 3D data for training, and also typically lack generalization beyond the training distribution. Moreover, they tend to be object centric and struggle with complex and intricate scenes. Conversely, 3D-free methods can generate text-controlled views of complex, in-the-wild scenes using a pretrained stable diffusion model without the need for a large amount of 3D-based training data, but lack camera control. In this paper, we introduce a method capable of generating camera-controlled viewpoints from a single input image, by combining the benefits of 3D-free and 3D-based approaches. Our method excels in handling complex and diverse scenes without extensive training or additional 3D and multiview data. It leverages widely available pretrained NVS models for weak guidance, integrating this knowledge into a 3D-free view synthesis style approach, along with enriching the CLIP vision-language space with 3D camera angle information, to achieve the desired results. Experimental results demonstrate that our method outperforms existing models in both qualitative and quantitative evaluations, achieving high-fidelity, consistent novel view synthesis at desired camera angles across a wide variety of scenes while maintaining accurate, natural detail representation and image clarity across various viewpoints. We also support our method with a comprehensive analysis of 2D image generation models and the 3D space, providing a solid foundation and rationale for our solution., Comment: 18 pages, 13 figures, v4: methodology and result improvement

We present TOPGN, a novel method for real-time transparent obstacle detection for robot navigation in unknown environments. We use a multi-layer 2D grid map representation obtained by summing the intensities of lidar point clouds that lie in multiple non-overlapping height intervals. We isolate a neighborhood of points reflected from transparent obstacles by comparing the intensities in the different 2D grid map layers. Using the neighborhood, we linearly extrapolate the transparent obstacle by computing a tangential line segment and use it to perform safe, real-time collision avoidance. Finally, we also demonstrate our transparent object isolation's applicability to mapping an environment. We demonstrate that our approach detects transparent objects made of various materials (glass, acrylic, PVC), arbitrary shapes, colors, and textures in a variety of real-world indoor and outdoor scenarios with varying lighting conditions. We compare our method with other glass/transparent object detection methods that use RGB images, 2D laser scans, etc. in these benchmark scenarios. We demonstrate superior detection accuracy in terms of F-score improvement at least by 12.74% and 38.46% decrease in mean absolute error (MAE), improved navigation success rates (at least two times better than the second-best), and a real-time inference rate (~50Hz on a mobile CPU). We will release our code and challenging benchmarks for future evaluations upon publication.

We present a multi-modal trajectory generation and selection algorithm for real-world mapless outdoor navigation in challenging scenarios with unstructured off-road features like buildings, grass, and curbs. Our goal is to compute suitable trajectories that (1) satisfy the environment-specific traversability constraints and (2) generate human-like paths while navigating in crosswalks, sidewalks, etc. Our formulation uses a Conditional Variational Autoencoder (CVAE) generative model enhanced with traversability constraints to generate multiple candidate trajectories for global navigation. We use VLMs and a visual prompting approach with their zero-shot ability of semantic understanding and logical reasoning to choose the best trajectory given the contextual information about the task. We evaluate our methods in various outdoor scenes with wheeled robots and compare the performance with other global navigation algorithms. In practice, we observe at least 3.35% improvement in traversability and 20.61% improvement in terms of human-like navigation in generated trajectories in challenging outdoor navigation scenarios.

We propose a new method for improving zero-shot ObjectNav that aims to utilize potentially available environmental percepts for navigational assistance. Our approach takes into account that the ground agent may have limited and sometimes obstructed view. Our formulation encourages Generative Communication (GC) between an assistive overhead agent with a global view containing the target object and the ground agent with an obfuscated view; both equipped with Vision-Language Models (VLMs) for vision-to-language translation. In this assisted setup, the embodied agents communicate environmental information before the ground agent executes actions towards a target. Despite the overhead agent having a global view with the target, we note a drop in performance (-13% in OSR and -13% in SPL) of a fully cooperative assistance scheme over an unassisted baseline. In contrast, a selective assistance scheme where the ground agent retains its independent exploratory behaviour shows a 10% OSR and 7.65% SPL improvement. To explain navigation performance, we analyze the GC for unique traits, quantifying the presence of hallucination and cooperation. Specifically, we identify the novel linguistic trait of preemptive hallucination in our embodied setting, where the overhead agent assumes that the ground agent has executed an action in the dialogue when it is yet to move, and note its strong correlation with navigation performance. We conduct real-world experiments and present some qualitative examples where we mitigate hallucinations via prompt finetuning to improve ObjectNav performance.

We present a new algorithm, Cross-Source-Context Place Recognition (CSCPR), for RGB-D indoor place recognition that integrates global retrieval and reranking into a single end-to-end model. Unlike prior approaches that primarily focus on the RGB domain, CSCPR is designed to handle the RGB-D data. We extend the Context-of-Clusters (CoCs) for handling noisy colorized point clouds and introduce two novel modules for reranking: the Self-Context Cluster (SCC) and Cross Source Context Cluster (CSCC), which enhance feature representation and match query-database pairs based on local features, respectively. We also present two new datasets, ScanNetIPR and ARKitIPR. Our experiments demonstrate that CSCPR significantly outperforms state-of-the-art models on these datasets by at least 36.5% in Recall@1 at ScanNet-PR dataset and 44% in new datasets. Code and datasets will be released.

Leveraging Large Language Models' remarkable proficiency in text-based tasks, recent works on Multi-modal LLMs (MLLMs) extend them to other modalities like vision and audio. However, the progress in these directions has been mostly focused on tasks that only require a coarse-grained understanding of the audio-visual semantics. We present Meerkat, an audio-visual LLM equipped with a fine-grained understanding of image and audio both spatially and temporally. With a new modality alignment module based on optimal transport and a cross-attention module that enforces audio-visual consistency, Meerkat can tackle challenging tasks such as audio referred image grounding, image guided audio temporal localization, and audio-visual fact-checking. Moreover, we carefully curate a large dataset AVFIT that comprises 3M instruction tuning samples collected from open-source datasets, and introduce MeerkatBench that unifies five challenging audio-visual tasks. We achieve state-of-the-art performance on all these downstream tasks with a relative improvement of up to 37.12%., Comment: Accepted at ECCV 2024

We present a multimodal learning-based method to simultaneously synthesize co-speech facial expressions and upper-body gestures for digital characters using RGB video data captured using commodity cameras. Our approach learns from sparse face landmarks and upper-body joints, estimated directly from video data, to generate plausible emotive character motions. Given a speech audio waveform and a token sequence of the speaker's face landmark motion and body-joint motion computed from a video, our method synthesizes the motion sequences for the speaker's face landmarks and body joints to match the content and the affect of the speech. We design a generator consisting of a set of encoders to transform all the inputs into a multimodal embedding space capturing their correlations, followed by a pair of decoders to synthesize the desired face and pose motions. To enhance the plausibility of synthesis, we use an adversarial discriminator that learns to differentiate between the face and pose motions computed from the original videos and our synthesized motions based on their affective expressions. To evaluate our approach, we extend the TED Gesture Dataset to include view-normalized, co-speech face landmarks in addition to body gestures. We demonstrate the performance of our method through thorough quantitative and qualitative experiments on multiple evaluation metrics and via a user study. We observe that our method results in low reconstruction error and produces synthesized samples with diverse facial expressions and body gestures for digital characters., Comment: 14 pages, 7 figures, 2 tables

Image-text contrastive models such as CLIP learn transferable and robust representations for zero-shot transfer to a variety of downstream tasks. However, to obtain strong downstream performances, prompts need to be carefully curated, which can be a tedious engineering task. To address the issue of manual prompt engineering, prompt-tuning is used where a set of contextual vectors are learned by leveraging information from the training data. Despite their effectiveness, existing prompt-tuning frameworks often lack interpretability, thus limiting their ability to understand the compositional nature of images. In this work, we first identify that incorporating compositional attributes (e.g., a "green" tree frog) in the design of manual prompts can significantly enhance image-text alignment scores. Building upon this observation, we propose a novel and interpretable prompt-tuning method named IntCoOp, which learns to jointly align attribute-level inductive biases and class embeddings during prompt-tuning. To assess the effectiveness of our approach, we evaluate IntCoOp across two representative tasks in a few-shot learning setup: generalization to novel classes, and unseen domain shifts. Through extensive experiments across 10 downstream datasets on CLIP, we find that introducing attribute-level inductive biases leads to superior performance against state-of-the-art prompt tuning frameworks. Notably, in a 16-shot setup, IntCoOp improves CoOp by 7.35% in average performance across 10 diverse datasets.

Perceiving and understanding non-speech sounds and non-verbal speech is essential to making decisions that help us interact with our surroundings. In this paper, we propose GAMA, a novel General-purpose Large Audio-Language Model (LALM) with Advanced Audio Understanding and Complex Reasoning Abilities. We build GAMA by integrating an LLM with multiple types of audio representations, including features from a custom Audio Q-Former, a multi-layer aggregator that aggregates features from multiple layers of an audio encoder. We fine-tune GAMA on a large-scale audio-language dataset, which augments it with audio understanding capabilities. Next, we propose CompA-R (Instruction-Tuning for Complex Audio Reasoning), a synthetically generated instruction-tuning (IT) dataset with instructions that require the model to perform complex reasoning on the input audio. We instruction-tune GAMA with CompA-R to endow it with complex reasoning abilities, where we further add a soft prompt as input with high-level semantic evidence by leveraging event tags of the input audio. Finally, we also propose CompA-R-test, a human-labeled evaluation dataset for evaluating the capabilities of LALMs on open-ended audio question-answering that requires complex reasoning. Through automated and expert human evaluations, we show that GAMA outperforms all other LALMs in literature on diverse audio understanding tasks by margins of 1%-84%. Further, GAMA IT-ed on CompA-R proves to be superior in its complex reasoning and instruction following capabilities., Comment: Project Website: https://sreyan88.github.io/gamaaudio/

Large language models (LLMs) have grown in popularity due to their natural language interface and pre trained knowledge, leading to rapidly increasing success in question-answering (QA) tasks. More recently, multi-agent systems with LLM-based agents (Multi-LLM) have been utilized increasingly more for QA. In these scenarios, the models may each answer the question and reach a consensus or each model is specialized to answer different domain questions. However, most prior work dealing with Multi-LLM QA has focused on scenarios where the models are asked in a zero-shot manner or are given information sources to extract the answer. For question answering of an unknown environment, embodied exploration of the environment is first needed to answer the question. This skill is necessary for personalizing embodied AI to environments such as households. There is a lack of insight into whether a Multi-LLM system can handle question-answering based on observations from embodied exploration. In this work, we address this gap by investigating the use of Multi-Embodied LLM Explorers (MELE) for QA in an unknown environment. Multiple LLM-based agents independently explore and then answer queries about a household environment. We analyze different aggregation methods to generate a single, final answer for each query: debating, majority voting, and training a central answer module (CAM). Using CAM, we observe a $46\%$ higher accuracy compared against the other non-learning-based aggregation methods. We provide code and the query dataset for further research., Comment: 16 pages, 9 Figures, 5 Tables

Large vision-language models (LVLMs) are prone to hallucinations, where certain contextual cues in an image can trigger the language module to produce overconfident and incorrect reasoning about abnormal or hypothetical objects. While some benchmarks have been developed to investigate LVLM hallucinations, they often rely on hand-crafted corner cases whose failure patterns may not generalize well. Additionally, fine-tuning on these examples could undermine their validity. To address this, we aim to scale up the number of cases through an automated approach, reducing human bias in crafting such corner cases. This motivates the development of AutoHallusion, the first automated benchmark generation approach that employs several key strategies to create a diverse range of hallucination examples. Our generated visual-question pairs pose significant challenges to LVLMs, requiring them to overcome contextual biases and distractions to arrive at correct answers. AutoHallusion enables us to create new benchmarks at the minimum cost and thus overcomes the fragility of hand-crafted benchmarks. It also reveals common failure patterns and reasons, providing key insights to detect, avoid, or control hallucinations. Comprehensive evaluations of top-tier LVLMs, e.g., GPT-4V(ision), Gemini Pro Vision, Claude 3, and LLaVA-1.5, show a 97.7% and 98.7% success rate of hallucination induction on synthetic and real-world datasets of AutoHallusion, paving the way for a long battle against hallucinations. The codebase and data can be accessed at https://github.com/wuxiyang1996/AutoHallusion.

Music is a universal language that can communicate emotions and feelings. It forms an essential part of the whole spectrum of creative media, ranging from movies to social media posts. Machine learning models that can synthesize music are predominantly conditioned on textual descriptions of it. Inspired by how musicians compose music not just from a movie script, but also through visualizations, we propose MeLFusion, a model that can effectively use cues from a textual description and the corresponding image to synthesize music. MeLFusion is a text-to-music diffusion model with a novel "visual synapse", which effectively infuses the semantics from the visual modality into the generated music. To facilitate research in this area, we introduce a new dataset MeLBench, and propose a new evaluation metric IMSM. Our exhaustive experimental evaluation suggests that adding visual information to the music synthesis pipeline significantly improves the quality of generated music, measured both objectively and subjectively, with a relative gain of up to 67.98% on the FAD score. We hope that our work will gather attention to this pragmatic, yet relatively under-explored research area., Comment: Accepted at CVPR 2024 as Highlight paper. Webpage: https://schowdhury671.github.io/melfusion_cvpr2024/

Visual cues, like lip motion, have been shown to improve the performance of Automatic Speech Recognition (ASR) systems in noisy environments. We propose LipGER (Lip Motion aided Generative Error Correction), a novel framework for leveraging visual cues for noise-robust ASR. Instead of learning the cross-modal correlation between the audio and visual modalities, we make an LLM learn the task of visually-conditioned (generative) ASR error correction. Specifically, we instruct an LLM to predict the transcription from the N-best hypotheses generated using ASR beam-search. This is further conditioned on lip motions. This approach addresses key challenges in traditional AVSR learning, such as the lack of large-scale paired datasets and difficulties in adapting to new domains. We experiment on 4 datasets in various settings and show that LipGER improves the Word Error Rate in the range of 1.1%-49.2%. We also release LipHyp, a large-scale dataset with hypothesis-transcription pairs that is additionally equipped with lip motion cues to promote further research in this space, Comment: InterSpeech 2024. Code and Data: https://github.com/Sreyan88/LipGER

We present ABEX, a novel and effective generative data augmentation methodology for low-resource Natural Language Understanding (NLU) tasks. ABEX is based on ABstract-and-EXpand, a novel paradigm for generating diverse forms of an input document -- we first convert a document into its concise, abstract description and then generate new documents based on expanding the resultant abstraction. To learn the task of expanding abstract descriptions, we first train BART on a large-scale synthetic dataset with abstract-document pairs. Next, to generate abstract descriptions for a document, we propose a simple, controllable, and training-free method based on editing AMR graphs. ABEX brings the best of both worlds: by expanding from abstract representations, it preserves the original semantic properties of the documents, like style and meaning, thereby maintaining alignment with the original label and data distribution. At the same time, the fundamental process of elaborating on abstract descriptions facilitates diverse generations. We demonstrate the effectiveness of ABEX on 4 NLU tasks spanning 12 datasets and 4 low-resource settings. ABEX outperforms all our baselines qualitatively with improvements of 0.04% - 38.8%. Qualitatively, ABEX outperforms all prior methods from literature in terms of context and length diversity., Comment: ACL 2024 Main Conference. Code and data: https://github.com/Sreyan88/ABEX

Aligning foundation models is essential for their safe and trustworthy deployment. However, traditional fine-tuning methods are computationally intensive and require updating billions of model parameters. A promising alternative, alignment via decoding, adjusts the response distribution directly without model updates to maximize a target reward $r$, thus providing a lightweight and adaptable framework for alignment. However, principled decoding methods rely on oracle access to an optimal Q-function ($Q^*$), which is often unavailable in practice. Hence, prior SoTA methods either approximate this $Q^*$ using $Q^{\pi_{\texttt{sft}}}$ (derived from the reference $\texttt{SFT}$ model) or rely on short-term rewards, resulting in sub-optimal decoding performance. In this work, we propose Transfer $Q^*$, which implicitly estimates the optimal value function for a target reward $r$ through a baseline model $\rho_{\texttt{BL}}$ aligned with a baseline reward $\rho_{\texttt{BL}}$ (which can be different from the target reward $r$). Theoretical analyses of Transfer $Q^*$ provide a rigorous characterization of its optimality, deriving an upper bound on the sub-optimality gap and identifying a hyperparameter to control the deviation from the pre-trained reference $\texttt{SFT}$ model based on user needs. Our approach significantly reduces the sub-optimality gap observed in prior SoTA methods and demonstrates superior empirical performance across key metrics such as coherence, diversity, and quality in extensive tests on several synthetic and real datasets.

We present a novel machine-learning (ML) approach (EM-GANSim) for real-time electromagnetic (EM) propagation that is used for wireless communication simulation in 3D indoor environments. Our approach uses a modified conditional Generative Adversarial Network (GAN) that incorporates encoded geometry and transmitter location while adhering to the electromagnetic propagation theory. The overall physically-inspired learning is able to predict the power distribution in 3D scenes, which is represented using heatmaps. Our overall accuracy is comparable to ray tracing-based EM simulation, as evidenced by lower mean squared error values. Furthermore, our GAN-based method drastically reduces the computation time, achieving a 5X speedup on complex benchmarks. In practice, it can compute the signal strength in a few milliseconds on any location in 3D indoor environments. We also present a large dataset of 3D models and EM ray tracing-simulated heatmaps. To the best of our knowledge, EM-GANSim is the first real-time algorithm for EM simulation in complex 3D indoor environments. We plan to release the code and the dataset., Comment: 10 pages, 8 figures, 5 tables

Better fuel efficiency leads to better financial security as well as a cleaner environment. We propose a novel approach for improving fuel efficiency in unstructured and unregulated traffic environments. Existing intelligent transportation solutions for improving fuel efficiency, however, apply only to traffic intersections with sparse traffic or traffic where drivers obey the regulations, or both. We propose GameOpt+, a novel hybrid approach for cooperative intersection control in dynamic, multi-lane, unsignalized intersections. GameOpt+ is a hybrid solution that combines an auction mechanism and an optimization-based trajectory planner. It generates a priority entrance sequence for each agent and computes velocity controls in real-time, taking less than 10 milliseconds even in high-density traffic with over 10,000 vehicles per hour. Compared to fully optimization-based methods, it operates 100 times faster while ensuring fairness, safety, and efficiency. Tested on the SUMO simulator, our algorithm improves throughput by at least 25%, reduces the time to reach the goal by at least 70%, and decreases fuel consumption by 50% compared to auction-based and signaled approaches using traffic lights and stop signs. GameOpt+ is also unaffected by unbalanced traffic inflows, whereas some of the other baselines encountered a decrease in performance in unbalanced traffic inflow environments., Comment: Journal Version

Large Vision-Language Models (LVLMs) often produce responses that misalign with factual information, a phenomenon known as hallucinations. While hallucinations are well-studied, the exact causes behind them remain underexplored. In this paper, we first investigate the root causes of hallucinations in LVLMs. Our findings reveal that existing mitigation techniques primarily reduce hallucinations for visual recognition prompts-those that require simple descriptions of visual elements-but fail for cognitive prompts that demand deliberate reasoning. We identify the core issue as a lack of true visual perception in LVLMs: although they can accurately recognize visual elements, they struggle to fully interpret these elements in the context of the input prompt and effectively link this recognition to their internal knowledge, which is critical for reasoning. To address this gap, we introduce Visual Description Grounded Decoding (VDGD), a simple, robust, and training-free method designed to enhance visual perception and improve reasoning capabilities in LVLMs. VDGD works by first generating a detailed description of the image and appending it as a prefix to the instruction. During response generation, tokens are sampled based on their KL divergence to the description, favoring candidates with lower divergence. Experimental results on multiple visual reasoning benchmarks and LVLMs demonstrate that VDGD consistently outperforms existing baselines 2% - 33%. Finally, we introduce VaLLu, a benchmark designed for comprehensive evaluation of the cognitive capabilities of LVLMs., Comment: Preprint. Under review

We present a method for multi-concept customization of pretrained text-to-video (T2V) models. Intuitively, the multi-concept customized video can be derived from the (non-linear) intersection of the video manifolds of the individual concepts, which is not straightforward to find. We hypothesize that sequential and controlled walking towards the intersection of the video manifolds, directed by text prompting, leads to the solution. To do so, we generate the various concepts and their corresponding interactions, sequentially, in an autoregressive manner. Our method can generate videos of multiple custom concepts (subjects, action and background) such as a teddy bear running towards a brown teapot, a dog playing violin and a teddy bear swimming in the ocean. We quantitatively evaluate our method using videoCLIP and DINO scores, in addition to human evaluation. Videos for results presented in this paper can be found at https://github.com/divyakraman/MultiConceptVideo2024., Comment: Paper accepted to AI4CC Workshop at CVPR 2024

We introduce a novel approach for prompt mixing, aiming to generate images at the intersection of multiple text prompts using pre-trained text-to-image diffusion models. At each time step during diffusion denoising, our algorithm forecasts predictions w.r.t. the generated image and makes informed text conditioning decisions. To do so, we leverage the connection between diffusion models (rooted in non-equilibrium thermodynamics) and the Black-Scholes model for pricing options in Finance, and draw analogies between the variables in both contexts to derive an appropriate algorithm for prompt mixing using the Black Scholes model. Specifically, the parallels between diffusion models and the Black-Scholes model enable us to leverage properties related to the dynamics of the Markovian model derived in the Black-Scholes algorithm. Our prompt-mixing algorithm is data-efficient, meaning it does not need additional training. Furthermore, it operates without human intervention or hyperparameter tuning. We highlight the benefits of our approach by comparing it qualitatively and quantitatively to other prompt mixing techniques, including linear interpolation, alternating prompts, step-wise prompt switching, and CLIP-guided prompt selection across various scenarios such as single object per text prompt, multiple objects per text prompt and objects against backgrounds. Code is available at https://github.com/divyakraman/BlackScholesDiffusion2024.

In this paper, we present LOC-ZSON, a novel Language-driven Object-Centric image representation for object navigation task within complex scenes. We propose an object-centric image representation and corresponding losses for visual-language model (VLM) fine-tuning, which can handle complex object-level queries. In addition, we design a novel LLM-based augmentation and prompt templates for stability during training and zero-shot inference. We implement our method on Astro robot and deploy it in both simulated and real-world environments for zero-shot object navigation. We show that our proposed method can achieve an improvement of 1.38 - 13.38% in terms of text-to-image recall on different benchmark settings for the retrieval task. For object navigation, we show the benefit of our approach in simulation and real world, showing 5% and 16.67% improvement in terms of navigation success rate, respectively., Comment: Accepted to ICRA 2024

We present and tackle the problem of Embodied Question Answering (EQA) with Situational Queries (S-EQA) in a household environment. Unlike prior EQA work tackling simple queries that directly reference target objects and properties ("What is the color of the car?"), situational queries (such as "Is the house ready for sleeptime?") are more challenging requiring the agent to identify multiple objects (Doors: Closed, Lights: Off, etc.) and reach a consensus on their states for an answer. Towards this objective, we first introduce a novel Prompt-Generate-Evaluate (PGE) scheme that wraps around an LLM's output to create a dataset of unique situational queries and corresponding consensus object information. PGE maintains uniqueness among the generated queries, using semantic similarity via a feedback loop. We annotate the generated data for ground truth answers via a large scale user-study conducted on M-Turk, and with a high answerability rate of 97.26%, establish that LLMs are good at generating situational data. However, using the same LLM to answer the queries gives a low success rate of 46.2%; indicating that while LLMs are good at generating query data, they are poor at answering them. We use images from the VirtualHome simulator with the S-EQA queries establish an evaluation benchmark via Visual Question Answering (VQA). We report an improved accuracy of 15.31% while using queries framed from the generated object consensus for VQA over directly answering situational ones, indicating that such simplification is necessary for improved performance. To the best of our knowledge, this is the first work to introduce EQA in the context of situational queries that also uses a generative approach for query creation. We aim to foster research on improving the real-world usability of embodied agents in household environments through this work., Comment: 9 Pages

In this paper, we present a new approach to bridge the domain gap between synthetic and real-world data for unmanned aerial vehicle (UAV)-based perception. Our formulation is designed for dynamic scenes, consisting of small moving objects or human actions. We propose an extension of K-Planes Neural Radiance Field (NeRF), wherein our algorithm stores a set of tiered feature vectors. The tiered feature vectors are generated to effectively model conceptual information about a scene as well as an image decoder that transforms output feature maps into RGB images. Our technique leverages the information amongst both static and dynamic objects within a scene and is able to capture salient scene attributes of high altitude videos. We evaluate its performance on challenging datasets, including Okutama Action and UG2, and observe considerable improvement in accuracy over state of the art neural rendering methods., Comment: 8 pages, submitted to ICRA2025

Home robots intend to make their users lives easier. Our work assists in this goal by enabling robots to inform their users of dangerous or unsanitary anomalies in their home. Some examples of these anomalies include the user leaving their milk out, forgetting to turn off the stove, or leaving poison accessible to children. To move towards enabling home robots with these abilities, we have created a new dataset, which we call SafetyDetect. The SafetyDetect dataset consists of 1000 anomalous home scenes, each of which contains unsafe or unsanitary situations for an agent to detect. Our approach utilizes large language models (LLMs) alongside both a graph representation of the scene and the relationships between the objects in the scene. Our key insight is that this connected scene graph and the object relationships it encodes enables the LLM to better reason about the scene -- especially as it relates to detecting dangerous or unsanitary situations. Our most promising approach utilizes GPT-4 and pursues a categorization technique where object relations from the scene graph are classified as normal, dangerous, unsanitary, or dangerous for children. This method is able to correctly identify over 90% of anomalous scenarios in the SafetyDetect Dataset. Additionally, we conduct real world experiments on a ClearPath TurtleBot where we generate a scene graph from visuals of the real world scene, and run our approach with no modification. This setup resulted in little performance loss. The SafetyDetect Dataset and code will be released to the public upon this papers publication.

We present AGL-NET, a novel learning-based method for global localization using LiDAR point clouds and satellite maps. AGL-NET tackles two critical challenges: bridging the representation gap between image and points modalities for robust feature matching, and handling inherent scale discrepancies between global view and local view. To address these challenges, AGL-NET leverages a unified network architecture with a novel two-stage matching design. The first stage extracts informative neural features directly from raw sensor data and performs initial feature matching. The second stage refines this matching process by extracting informative skeleton features and incorporating a novel scale alignment step to rectify scale variations between LiDAR and map data. Furthermore, a novel scale and skeleton loss function guides the network toward learning scale-invariant feature representations, eliminating the need for pre-processing satellite maps. This significantly improves real-world applicability in scenarios with unknown map scales. To facilitate rigorous performance evaluation, we introduce a meticulously designed dataset within the CARLA simulator specifically tailored for metric localization training and assessment. The code and data can be accessed at https://github.com/rayguan97/AGL-Net.

We present a novel end-to-end algorithm (PoCo) for the indoor RGB-D place recognition task, aimed at identifying the most likely match for a given query frame within a reference database. The task presents inherent challenges attributed to the constrained field of view and limited range of perception sensors. We propose a new network architecture, which generalizes the recent Context of Clusters (CoCs) to extract global descriptors directly from the noisy point clouds through end-to-end learning. Moreover, we develop the architecture by integrating both color and geometric modalities into the point features to enhance the global descriptor representation. We conducted evaluations on public datasets ScanNet-PR and ARKit with 807 and 5047 scenarios, respectively. PoCo achieves SOTA performance: on ScanNet-PR, we achieve R@1 of 64.63%, a 5.7% improvement from the best-published result CGis (61.12%); on Arkit, we achieve R@1 of 45.12%, a 13.3% improvement from the best-published result CGis (39.82%). In addition, PoCo shows higher efficiency than CGis in inference time (1.75X-faster), and we demonstrate the effectiveness of PoCo in recognizing places within a real-world laboratory environment.

Open-vocabulary vision-language models (VLMs) like CLIP, trained using contrastive loss, have emerged as a promising new paradigm for text-to-image retrieval. However, do VLMs understand compound nouns (CNs) (e.g., lab coat) as well as they understand nouns (e.g., lab)? We curate Compun, a novel benchmark with 400 unique and commonly used CNs, to evaluate the effectiveness of VLMs in interpreting CNs. The Compun benchmark challenges a VLM for text-to-image retrieval where, given a text prompt with a CN, the task is to select the correct image that shows the CN among a pair of distractor images that show the constituent nouns that make up the CN. Next, we perform an in-depth analysis to highlight CLIPs' limited understanding of certain types of CNs. Finally, we present an alternative framework that moves beyond hand-written templates for text prompts widely used by CLIP-like models. We employ a Large Language Model to generate multiple diverse captions that include the CN as an object in the scene described by the caption. Our proposed method improves CN understanding of CLIP by 8.25% on Compun. Code and benchmark are available at: https://github.com/sonalkum/Compun, Comment: Accepted to NAACL 2024 Main Conference

We present CoDa (Constrained Generation based Data Augmentation), a controllable, effective, and training-free data augmentation technique for low-resource (data-scarce) NLP. Our approach is based on prompting off-the-shelf instruction-following Large Language Models (LLMs) for generating text that satisfies a set of constraints. Precisely, we extract a set of simple constraints from every instance in the low-resource dataset and verbalize them to prompt an LLM to generate novel and diverse training instances. Our findings reveal that synthetic data that follows simple constraints in the downstream dataset act as highly effective augmentations, and CoDa can achieve this without intricate decoding-time constrained generation techniques or fine-tuning with complex algorithms that eventually make the model biased toward the small number of training instances. Additionally, CoDa is the first framework that provides users explicit control over the augmentation generation process, thereby also allowing easy adaptation to several domains. We demonstrate the effectiveness of CoDa across 11 datasets spanning 3 tasks and 3 low-resource settings. CoDa outperforms all our baselines, qualitatively and quantitatively, with improvements of 0.12%-7.19%. Code is available here: https://github.com/Sreyan88/CoDa, Comment: Accepted to NAACL 2024 Findings

We propose VLM-Social-Nav, a novel Vision-Language Model (VLM) based navigation approach to compute a robot's motion in human-centered environments. Our goal is to make real-time decisions on robot actions that are socially compliant with human expectations. We utilize a perception model to detect important social entities and prompt a VLM to generate guidance for socially compliant robot behavior. VLM-Social-Nav uses a VLM-based scoring module that computes a cost term that ensures socially appropriate and effective robot actions generated by the underlying planner. Our overall approach reduces reliance on large training datasets and enhances adaptability in decision-making. In practice, it results in improved socially compliant navigation in human-shared environments. We demonstrate and evaluate our system in four different real-world social navigation scenarios with a Turtlebot robot. We observe at least 27.38% improvement in the average success rate and 19.05% improvement in the average collision rate in the four social navigation scenarios. Our user study score shows that VLM-Social-Nav generates the most socially compliant navigation behavior.

We present ConVOI, a novel method for autonomous robot navigation in real-world indoor and outdoor environments using Vision Language Models (VLMs). We employ VLMs in two ways: first, we leverage their zero-shot image classification capability to identify the context or scenario (e.g., indoor corridor, outdoor terrain, crosswalk, etc) of the robot's surroundings, and formulate context-based navigation behaviors as simple text prompts (e.g. ``stay on the pavement"). Second, we utilize their state-of-the-art semantic understanding and logical reasoning capabilities to compute a suitable trajectory given the identified context. To this end, we propose a novel multi-modal visual marking approach to annotate the obstacle-free regions in the RGB image used as input to the VLM with numbers, by correlating it with a local occupancy map of the environment. The marked numbers ground image locations in the real-world, direct the VLM's attention solely to navigable locations, and elucidate the spatial relationships between them and terrains depicted in the image to the VLM. Next, we query the VLM to select numbers on the marked image that satisfy the context-based behavior text prompt, and construct a reference path using the selected numbers. Finally, we propose a method to extrapolate the reference trajectory when the robot's environmental context has not changed to prevent unnecessary VLM queries. We use the reference trajectory to guide a motion planner, and demonstrate that it leads to human-like behaviors (e.g. not cutting through a group of people, using crosswalks, etc.) in various real-world indoor and outdoor scenarios., Comment: 9 pages, 4 figures