Search

The rise of large language models (LLMs) and their tight integration into our daily life make it essential to dedicate efforts towards their trustworthiness. Uncertainty quantification for LLMs can establish more human trust into their responses, but also allows LLM agents to make more informed decisions based on each other's uncertainty. To estimate the uncertainty in a response, internal token logits, task-specific proxy models, or sampling of multiple responses are commonly used. This work focuses on asking the LLM itself to verbalize its uncertainty with a confidence score as part of its output tokens, which is a promising way for prompt- and model-agnostic uncertainty quantification with low overhead. Using an extensive benchmark, we assess the reliability of verbalized confidence scores with respect to different datasets, models, and prompt methods. Our results reveal that the reliability of these scores strongly depends on how the model is asked, but also that it is possible to extract well-calibrated confidence scores with certain prompt methods. We argue that verbalized confidence scores can become a simple but effective and versatile uncertainty quantification method in the future. Our code is available at https://github.com/danielyxyang/llm-verbalized-uq .

This study presents the first implementation of multilayer neural networks on a memristor/CMOS integrated system on chip (SoC) to simultaneously detect multiple diseases. To overcome limitations in medical data, generative AI techniques are used to enhance the dataset, improving the classifier's robustness and diversity. The system achieves notable performance with low latency, high accuracy (91.82%), and energy efficiency, facilitated by end-to-end execution on a memristor-based SoC with ten 256x256 crossbar arrays and an integrated on-chip processor. This research showcases the transformative potential of memristive in-memory computing hardware in accelerating machine learning applications for medical diagnostics., Comment: 14 pages, 5 figures

We introduce SeaSplat, a method to enable real-time rendering of underwater scenes leveraging recent advances in 3D radiance fields. Underwater scenes are challenging visual environments, as rendering through a medium such as water introduces both range and color dependent effects on image capture. We constrain 3D Gaussian Splatting (3DGS), a recent advance in radiance fields enabling rapid training and real-time rendering of full 3D scenes, with a physically grounded underwater image formation model. Applying SeaSplat to the real-world scenes from SeaThru-NeRF dataset, a scene collected by an underwater vehicle in the US Virgin Islands, and simulation-degraded real-world scenes, not only do we see increased quantitative performance on rendering novel viewpoints from the scene with the medium present, but are also able to recover the underlying true color of the scene and restore renders to be without the presence of the intervening medium. We show that the underwater image formation helps learn scene structure, with better depth maps, as well as show that our improvements maintain the significant computational improvements afforded by leveraging a 3D Gaussian representation., Comment: Project page here: https://seasplat.github.io

Teaching robots novel skills with demonstrations via human-in-the-loop data collection techniques like kinesthetic teaching or teleoperation puts a heavy burden on human supervisors. In contrast to this paradigm, it is often significantly easier to provide raw, action-free visual data of tasks being performed. Moreover, this data can even be mined from video datasets or the web. Ideally, this data can serve to guide robot learning for new tasks in novel environments, informing both "what" to do and "how" to do it. A powerful way to encode both the "what" and the "how" is to infer a well-shaped reward function for reinforcement learning. The challenge is determining how to ground visual demonstration inputs into a well-shaped and informative reward function. We propose a technique Rank2Reward for learning behaviors from videos of tasks being performed without access to any low-level states and actions. We do so by leveraging the videos to learn a reward function that measures incremental "progress" through a task by learning how to temporally rank the video frames in a demonstration. By inferring an appropriate ranking, the reward function is able to guide reinforcement learning by indicating when task progress is being made. This ranking function can be integrated into an adversarial imitation learning scheme resulting in an algorithm that can learn behaviors without exploiting the learned reward function. We demonstrate the effectiveness of Rank2Reward at learning behaviors from raw video on a number of tabletop manipulation tasks in both simulations and on a real-world robotic arm. We also demonstrate how Rank2Reward can be easily extended to be applicable to web-scale video datasets., Comment: ICRA 2024

Multi-modal learning has emerged as an increasingly promising avenue in vision recognition, driving innovations across diverse domains ranging from media and education to healthcare and transportation. Despite its success, the robustness of multi-modal learning for visual recognition is often challenged by the unavailability of a subset of modalities, especially the visual modality. Conventional approaches to mitigate missing modalities in multi-modal learning rely heavily on algorithms and modality fusion schemes. In contrast, this paper explores the use of text-to-image models to assist multi-modal learning. Specifically, we propose a simple but effective multi-modal learning framework GTI-MM to enhance the data efficiency and model robustness against missing visual modality by imputing the missing data with generative transformers. Using multiple multi-modal datasets with visual recognition tasks, we present a comprehensive analysis of diverse conditions involving missing visual modality in data, including model training. Our findings reveal that synthetic images benefit training data efficiency with visual data missing in training and improve model robustness with visual data missing involving training and testing. Moreover, we demonstrate GTI-MM is effective with lower generation quantity and simple prompt techniques.

The lack of contextual information in text data can make the annotation process of text-based emotion classification datasets challenging. As a result, such datasets often contain labels that fail to consider all the relevant emotions in the vocabulary. This misalignment between text inputs and labels can degrade the performance of machine learning models trained on top of them. As re-annotating entire datasets is a costly and time-consuming task that cannot be done at scale, we propose to use the expressive capabilities of large language models to synthesize additional context for input text to increase its alignment with the annotated emotional labels. In this work, we propose a formal definition of textual context to motivate a prompting strategy to enhance such contextual information. We provide both human and empirical evaluation to demonstrate the efficacy of the enhanced context. Our method improves alignment between inputs and their human-annotated labels from both an empirical and human-evaluated standpoint.

Video summarization remains a huge challenge in computer vision due to the size of the input videos to be summarized. We propose an efficient, language-only video summarizer that achieves competitive accuracy with high data efficiency. Using only textual captions obtained via a zero-shot approach, we train a language transformer model and forego image representations. This method allows us to perform filtration amongst the representative text vectors and condense the sequence. With our approach, we gain explainability with natural language that comes easily for human interpretation and textual summaries of the videos. An ablation study that focuses on modality and data compression shows that leveraging text modality only effectively reduces input data processing while retaining comparable results., Comment: \c{opyright} 2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

Coral reefs are fast-changing and complex ecosystems that are crucial to monitor and study. Biological hotspot detection can help coral reef managers prioritize limited resources for monitoring and intervention tasks. Here, we explore the use of autonomous underwater vehicles (AUVs) with cameras, coupled with visual detectors and photogrammetry, to map and identify these hotspots. This approach can provide high spatial resolution information in fast feedback cycles. To the best of our knowledge, we present one of the first attempts at using an AUV to gather visually-observed, fine-grain biological hotspot maps in concert with topography of a coral reefs. Our hotspot maps correlate with rugosity, an established proxy metric for coral reef biodiversity and abundance, as well as with our visual inspections of the 3D reconstruction. We also investigate issues of scaling this approach when applied to new reefs by using these visual detectors pre-trained on large public datasets., Comment: CV4Animals Workshop at CVPR 2023

We evaluated the temporal performance of a deep learning (DL) based artificial intelligence (AI) model for auto segmentation in prostate radiotherapy, seeking to correlate its efficacy with changes in clinical landscapes. Our study involved 1328 prostate cancer patients who underwent definitive radiotherapy from January 2006 to August 2022 at the University of Texas Southwestern Medical Center. We trained a UNet based segmentation model on data from 2006 to 2011 and tested it on data from 2012 to 2022 to simulate real world clinical deployment. We measured the model performance using the Dice similarity coefficient (DSC), visualized the trends in contour quality using exponentially weighted moving average (EMA) curves. Additionally, we performed Wilcoxon Rank Sum Test to analyze the differences in DSC distributions across distinct periods, and multiple linear regression to investigate the impact of various clinical factors. The model exhibited peak performance in the initial phase (from 2012 to 2014) for segmenting the prostate, rectum, and bladder. However, we observed a notable decline in performance for the prostate and rectum after 2015, while bladder contour quality remained stable. Key factors that impacted the prostate contour quality included physician contouring styles, the use of various hydrogel spacer, CT scan slice thickness, MRI-guided contouring, and using intravenous (IV) contrast. Rectum contour quality was influenced by factors such as slice thickness, physician contouring styles, and the use of various hydrogel spacers. The bladder contour quality was primarily affected by using IV contrast. This study highlights the challenges in maintaining AI model performance consistency in a dynamic clinical setting. It underscores the need for continuous monitoring and updating of AI models to ensure their ongoing effectiveness and relevance in patient care.

This paper presents a method for large-scale retrieval of piano sheet music images. Our work differs from previous studies on sheet music retrieval in two ways. First, we investigate the problem at a much larger scale than previous studies, using all solo piano sheet music images in the entire IMSLP dataset as a searchable database. Second, we use cell phone images of sheet music as our input queries, which lends itself to a practical, user-facing application. We show that a previously proposed fingerprinting method for sheet music retrieval is far too slow for a real-time application, and we diagnose its shortcomings. We propose a novel hashing scheme called dynamic n-gram fingerprinting that significantly reduces runtime while simultaneously boosting retrieval accuracy. In experiments on IMSLP data, our proposed method achieves a mean reciprocal rank of 0.85 and an average runtime of 0.98 seconds per query., Comment: 8 pages, 3 figures, 2 tables. Accepted paper at the International Society for Music Information Retrieval Conference (ISMIR) 2020

This article investigates a cross-modal retrieval problem in which a user would like to retrieve a passage of music from a MIDI file by taking a cell phone picture of several lines of sheet music. This problem is challenging for two reasons: it has a significant runtime constraint since it is a user-facing application, and there is very little relevant training data containing cell phone images of sheet music. To solve this problem, we introduce a novel feature representation called a bootleg score which encodes the position of noteheads relative to staff lines in sheet music. The MIDI representation can be converted into a bootleg score using deterministic rules of Western musical notation, and the sheet music image can be converted into a bootleg score using classical computer vision techniques for detecting simple geometrical shapes. Once the MIDI and cell phone image have been converted into bootleg scores, we can estimate the alignment using dynamic programming. The most notable characteristic of our system is that it has no trainable weights at all -- only a set of about 40 hyperparameters. With a training set of just 400 images, we show that our system generalizes well to a much larger set of 1600 test images from 160 unseen musical scores. Our system achieves a test F measure score of 0.89, has an average runtime of 0.90 seconds, and outperforms baseline systems based on music object detection and sheet-audio alignment. We provide extensive experimental validation and analysis of our system., Comment: 13 pages, 8 figures, 3 tables. Accepted article in IEEE Transactions on Multimedia. arXiv admin note: text overlap with arXiv:2004.10347

This paper investigates a cross-modal retrieval problem in which a user would like to retrieve a passage of music from a MIDI file by taking a cell phone picture of a physical page of sheet music. While audio-sheet music retrieval has been explored by a number of works, this scenario is novel in that the query is a cell phone picture rather than a digital scan. To solve this problem, we introduce a mid-level feature representation called a bootleg score which explicitly encodes the rules of Western musical notation. We convert both the MIDI and the sheet music into bootleg scores using deterministic rules of music and classical computer vision techniques for detecting simple geometric shapes. Once the MIDI and cell phone image have been converted into bootleg scores, we estimate the alignment using dynamic programming. The most notable characteristic of our system is that it does test-time adaptation and has no trainable weights at all -- only a set of about 30 hyperparameters. On a dataset containing 1000 cell phone pictures taken of 100 scores of classical piano music, our system achieves an F measure score of .869 and outperforms baseline systems based on commercial optical music recognition software., Comment: 8 pages, 8 figures, 1 table. Accepted paper at the International Society for Music Information Retrieval Conference (ISMIR) 2019

We present a novel approach to robotic grasp planning using both a learned grasp proposal network and a learned 3D shape reconstruction network. Our system generates 6-DOF grasps from a single RGB-D image of the target object, which is provided as input to both networks. By using the geometric reconstruction to refine the the candidate grasp produced by the grasp proposal network, our system is able to accurately grasp both known and unknown objects, even when the grasp location on the object is not visible in the input image. This paper presents the network architectures, training procedures, and grasp refinement method that comprise our system. Experiments demonstrate the efficacy of our system at grasping both known and unknown objects (91% success rate in a physical robot environment, 84% success rate in a simulated environment). We additionally perform ablation studies that show the benefits of combining a learned grasp proposal with geometric reconstruction for grasping, and also show that our system outperforms several baselines in a grasping task., Comment: 7 pages, 7 figures

A major challenge in reinforcement learning is exploration, when local dithering methods such as epsilon-greedy sampling are insufficient to solve a given task. Many recent methods have proposed to intrinsically motivate an agent to seek novel states, driving the agent to discover improved reward. However, while state-novelty exploration methods are suitable for tasks where novel observations correlate well with improved reward, they may not explore more efficiently than epsilon-greedy approaches in environments where the two are not well-correlated. In this paper, we distinguish between exploration tasks in which seeking novel states aids in finding new reward, and those where it does not, such as goal-conditioned tasks and escaping local reward maxima. We propose a new exploration objective, maximizing the reward prediction error (RPE) of a value function trained to predict extrinsic reward. We then propose a deep reinforcement learning method, QXplore, which exploits the temporal difference error of a Q-function to solve hard exploration tasks in high-dimensional MDPs. We demonstrate the exploration behavior of QXplore on several OpenAI Gym MuJoCo tasks and Atari games and observe that QXplore is comparable to or better than a baseline state-novelty method in all cases, outperforming the baseline on tasks where state novelty is not well-correlated with improved reward., Comment: Published at IJCAI 2020, camera-ready version

Treating children with autism spectrum disorders (ASD) with behavioral interventions, such as Pivotal Response Treatment (PRT), has shown promise in recent studies. However, deciding which therapy is best for a given patient is largely by trial and error, and choosing an ineffective intervention results in loss of valuable treatment time. We propose predicting patient response to PRT from baseline task-based fMRI by the novel application of a random forest and tree bagging strategy. Our proposed learning pipeline uses random forest regression to determine candidate brain voxels that may be informative in predicting treatment response. The candidate voxels are then tested stepwise for inclusion in a bagged tree ensemble. After the predictive model is constructed, bias correction is performed to further increase prediction accuracy. Using data from 19 ASD children who underwent a 16 week trial of PRT and a leave-one-out cross-validation framework, the presented learning pipeline was tested against several standard methods and variations of the pipeline and resulted in the highest prediction accuracy., Comment: Multimodal Learning for Clinical Decision Support (ML-CDS) 2016