Search

Analytical wake models provide a computationally efficient means to predict velocity distributions in wind turbine wakes in the atmospheric boundary layer (ABL). Most existing models are developed for neutral atmospheric conditions and correspondingly neglect the effects of buoyancy and Coriolis forces that lead to veer, i.e. changes in the wind direction with height. Both veer and changes in thermal stratification lead to lateral shearing of the wake behind a wind turbine, which affects the power output of downstream turbines. Here we develop an analytical engineering wake model for a wind turbine in yaw in ABL flows including Coriolis and thermal stratification effects. The model combines the new analytical representation of ABL vertical structure based on coupling Ekman and surface layer descriptions (Narasimhan, Gayme, and Meneveau, 2024a) with the vortex sheet-based wake model for yawed turbines (Bastankhah et al., 2022), as well as a new method to predict the wake expansion rate based on the Townsend-Perry logarithmic scaling of streamwise velocity variance. The proposed wake model's predictions show good agreement with Large Eddy Simulation (LES) results, capturing the effects of wind veer and yawing including the curled and sheared wake structures across various states of the ABL, ranging from neutrally to strongly stably stratified atmospheric conditions. The model significantly improves power loss predictions from wake interactions, especially in strongly stably stratified conditions where wind veer effects dominate.

In recent years, significant progress has been made in collecting large-scale datasets to improve segmentation and autonomous driving models. These large-scale datasets are often dominated by common environmental conditions such as "Clear and Day" weather, leading to decreased performance in under-represented conditions like "Rainy and Night". To address this issue, we introduce SynDiff-AD, a novel data augmentation pipeline that leverages diffusion models (DMs) to generate realistic images for such subgroups. SynDiff-AD uses ControlNet-a DM that guides data generation conditioned on semantic maps-along with a novel prompting scheme that generates subgroup-specific, semantically dense prompts. By augmenting datasets with SynDiff-AD, we improve the performance of segmentation models like Mask2Former and SegFormer by up to 1.2% and 2.3% on the Waymo dataset, and up to 1.4% and 0.7% on the DeepDrive dataset, respectively. Additionally, we demonstrate that our SynDiff-AD pipeline enhances the driving performance of end-to-end autonomous driving models, like AIM-2D and AIM-BEV, by up to 20% across diverse environmental conditions in the CARLA autonomous driving simulator, providing a more robust model., Comment: 15 pages, 10 figures

Automatic Speech Translation (AST) datasets for Indian languages remain critically scarce, with public resources covering fewer than 10 of the 22 official languages. This scarcity has resulted in AST systems for Indian languages lagging far behind those available for high-resource languages like English. In this paper, we first evaluate the performance of widely-used AST systems on Indian languages, identifying notable performance gaps and challenges. Our findings show that while these systems perform adequately on read speech, they struggle significantly with spontaneous speech, including disfluencies like pauses and hesitations. Additionally, there is a striking absence of systems capable of accurately translating colloquial and informal language, a key aspect of everyday communication. To this end, we introduce BhasaAnuvaad, the largest publicly available dataset for AST involving 13 out of 22 scheduled Indian languages and English spanning over 44,400 hours and 17M text segments. BhasaAnuvaad contains data for English speech to Indic text, as well as Indic speech to English text. This dataset comprises three key categories: (1) Curated datasets from existing resources, (2) Large-scale web mining, and (3) Synthetic data generation. By offering this diverse and expansive dataset, we aim to bridge the resource gap and promote advancements in AST for Indian languages., Comment: Work in Progress

Low-Rank Adaptation (LoRA) is a popular technique for parameter-efficient fine-tuning of Large Language Models (LLMs). We study how different LoRA modules can be merged to achieve skill composition -- testing the performance of the merged model on a target task that involves combining multiple skills, each skill coming from a single LoRA. This setup is favorable when it is difficult to obtain training data for the target task and when it can be decomposed into multiple skills. First, we identify practically occurring use-cases that can be studied under the realm of skill composition, e.g. solving hard math-word problems with code, creating a bot to answer questions on proprietary manuals or about domain-specialized corpora. Our main contribution is to show that concatenation of LoRAs (CAT), which optimally weights LoRAs that were individually trained on different skills, outperforms existing model- and data- merging techniques; for instance on math-word problems, CAT beats these methods by an average of 43% and 12% respectively. Thus, this paper advocates model merging as an efficient way to solve compositional tasks and underscores CAT as a simple, compute-friendly and effective procedure. To our knowledge, this is the first work demonstrating the superiority of model merging over data mixing for binary skill composition tasks. Code and data are available at https://github.com/aksh555/LoRA-Soups, Comment: COLING 2025 Industry track; 9 pages plus references and appendices

Time series forecasts are often influenced by exogenous contextual features in addition to their corresponding history. For example, in financial settings, it is hard to accurately predict a stock price without considering public sentiments and policy decisions in the form of news articles, tweets, etc. Though this is common knowledge, the current state-of-the-art (SOTA) forecasting models fail to incorporate such contextual information, owing to its heterogeneity and multimodal nature. To address this, we introduce ContextFormer, a novel plug-and-play method to surgically integrate multimodal contextual information into existing pre-trained forecasting models. ContextFormer effectively distills forecast-specific information from rich multimodal contexts, including categorical, continuous, time-varying, and even textual information, to significantly enhance the performance of existing base forecasters. ContextFormer outperforms SOTA forecasting models by up to 30% on a range of real-world datasets spanning energy, traffic, environmental, and financial domains.

Generating realistic time series samples is crucial for stress-testing models and protecting user privacy by using synthetic data. In engineering and safety-critical applications, these samples must meet certain hard constraints that are domain-specific or naturally imposed by physics or nature. Consider, for example, generating electricity demand patterns with constraints on peak demand times. This can be used to stress-test the functioning of power grids during adverse weather conditions. Existing approaches for generating constrained time series are either not scalable or degrade sample quality. To address these challenges, we introduce Constrained Posterior Sampling (CPS), a diffusion-based sampling algorithm that aims to project the posterior mean estimate into the constraint set after each denoising update. Notably, CPS scales to a large number of constraints (~100) without requiring additional training. We provide theoretical justifications highlighting the impact of our projection step on sampling. Empirically, CPS outperforms state-of-the-art methods in sample quality and similarity to real time series by around 10% and 42%, respectively, on real-world stocks, traffic, and air quality datasets.

As premodern texts are passed down over centuries, errors inevitably accrue. These errors can be challenging to identify, as some have survived undetected for so long precisely because they are so elusive. While prior work has evaluated error detection methods on artificially-generated errors, we introduce the first dataset of real errors in premodern Greek, enabling the evaluation of error detection methods on errors that genuinely accumulated at some stage in the centuries-long copying process. To create this dataset, we use metrics derived from BERT conditionals to sample 1,000 words more likely to contain errors, which are then annotated and labeled by a domain expert as errors or not. We then propose and evaluate new error detection methods and find that our discriminator-based detector outperforms all other methods, improving the true positive rate for classifying real errors by 5%. We additionally observe that scribal errors are more difficult to detect than print or digitization errors. Our dataset enables the evaluation of error detection methods on real errors in premodern texts for the first time, providing a benchmark for developing more effective error detection algorithms to assist scholars in restoring premodern works.

Assume that a target is known to be present at an unknown point among a finite set of locations in the plane. We search for it using a mobile robot that has imperfect sensing capabilities. It takes time for the robot to move between locations and search a location; we have a total time budget within which to conduct the search. We study the problem of computing a search path/strategy for the robot that maximizes the probability of detection of the target. Considering non-uniform travel times between points (e.g., based on the distance between them) is crucial for search and rescue applications; such problems have been investigated to a limited extent due to their inherent complexity. In this paper, we describe fast algorithms with performance guarantees for this search problem and some variants, complement them with complexity results, and perform experiments to observe their performance., Comment: 10 pages, 6 figures, 3 algorithms

Autonomous systems for software engineering are now capable of fixing bugs and developing features. These systems are commonly evaluated on SWE-bench (Jimenez et al., 2024a), which assesses their ability to solve software issues from GitHub repositories. However, SWE-bench uses only Python repositories, with problem statements presented predominantly as text and lacking visual elements such as images. This limited coverage motivates our inquiry into how existing systems might perform on unrepresented software engineering domains (e.g., front-end, game development, DevOps), which use different programming languages and paradigms. Therefore, we propose SWE-bench Multimodal (SWE-bench M), to evaluate systems on their ability to fix bugs in visual, user-facing JavaScript software. SWE-bench M features 617 task instances collected from 17 JavaScript libraries used for web interface design, diagramming, data visualization, syntax highlighting, and interactive mapping. Each SWE-bench M task instance contains at least one image in its problem statement or unit tests. Our analysis finds that top-performing SWE-bench systems struggle with SWE-bench M, revealing limitations in visual problem-solving and cross-language generalization. Lastly, we show that SWE-agent's flexible language-agnostic features enable it to substantially outperform alternatives on SWE-bench M, resolving 12% of task instances compared to 6% for the next best system.

Although language model (LM) agents are demonstrating growing potential in many domains, their success in cybersecurity has been limited due to simplistic design and the lack of fundamental features for this domain. We present EnIGMA, an LM agent for autonomously solving Capture The Flag (CTF) challenges. EnIGMA introduces new Agent-Computer Interfaces (ACIs) to improve the success rate on CTF challenges. We establish the novel Interactive Agent Tool concept, which enables LM agents to run interactive command-line utilities essential for these challenges. Empirical analysis of EnIGMA on over 350 CTF challenges from three different benchmarks indicates that providing a robust set of new tools with demonstration of their usage helps the LM solve complex problems and achieves state-of-the-art results on the NYU CTF and Intercode-CTF benchmarks. Finally, we discuss insights on ACI design and agent behavior on cybersecurity tasks that highlight the need to adapt real-world tools for LM agents.

Extreme events are of great importance since they often represent impactive occurrences. For instance, in terms of climate and weather, extreme events might be major storms, floods, extreme heat or cold waves, and more. However, they are often located at the tail of the data distribution. Consequently, accurately predicting these extreme events is challenging due to their rarity and irregularity. Prior studies have also referred to this as the out-of-distribution (OOD) problem, which occurs when the distribution of the test data is substantially different from that used for training. In this work, we propose two strategies, reweighting and fine-tuning, to tackle the challenge. Reweighting is a strategy used to force machine learning models to focus on extreme events, which is achieved by a weighted loss function that assigns greater penalties to the prediction errors for the extreme samples relative to those on the remainder of the data. Unlike previous intuitive reweighting methods based on simple heuristics of data distribution, we employ meta-learning to dynamically optimize these penalty weights. To further boost the performance on extreme samples, we start from the reweighted models and fine-tune them using only rare extreme samples. Through extensive experiments on multiple data sets, we empirically validate that our meta-learning-based reweighting outperforms existing heuristic ones, and the fine-tuning strategy can further increase the model performance. More importantly, these two strategies are model-agnostic, which can be implemented on any type of neural network for time series forecasting. The open-sourced code is available at \url{https://github.com/JimengShi/ReFine}.

Service robots in human-centered environments such as hospitals, office buildings, and long-term care homes need to navigate while adhering to social norms to ensure the safety and comfortability of the people they are sharing the space with. Furthermore, they need to adapt to new social scenarios that can arise during robot navigation. In this paper, we present a novel Online Lifelong Vision Language architecture, OLiVia-Nav, which uniquely integrates vision-language models (VLMs) with an online lifelong learning framework for robot social navigation. We introduce a unique distillation approach, Social Context Contrastive Language Image Pre-training (SC-CLIP), to transfer the social reasoning capabilities of large VLMs to a lightweight VLM, in order for OLiVia-Nav to directly encode social and environment context during robot navigation. These encoded embeddings are used to generate and select robot social compliant trajectories. The lifelong learning capabilities of SC-CLIP enable OLiVia-Nav to update the lightweight VLM with robot trajectory predictions overtime as new social scenarios are encountered. We conducted extensive real-world experiments in diverse social navigation scenarios. The results showed that OLiVia-Nav outperformed existing state-of-the-art DRL and VLM methods in terms of mean squared error, Hausdorff loss, and personal space violation duration. Ablation studies also verified the design choices for OLiVia-Nav.

Accurate weather forecasting is critical for science and society. Yet, existing methods have not managed to simultaneously have the properties of high accuracy, low uncertainty, and high computational efficiency. On one hand, to quantify the uncertainty in weather predictions, the strategy of ensemble forecast (i.e., generating a set of diverse predictions) is often employed. However, traditional ensemble numerical weather prediction (NWP) is computationally intensive. On the other hand, most existing machine learning-based weather prediction (MLWP) approaches are efficient and accurate. Nevertheless, they are deterministic and cannot capture the uncertainty of weather forecasting. In this work, we propose CoDiCast, a conditional diffusion model to generate accurate global weather prediction, while achieving uncertainty quantification with ensemble forecasts and modest computational cost. The key idea is to simulate a conditional version of the reverse denoising process in diffusion models, which starts from pure Gaussian noise to generate realistic weather scenarios for a future time point. Each denoising step is conditioned on observations from the recent past. Ensemble forecasts are achieved by repeatedly sampling from stochastic Gaussian noise to represent uncertainty quantification. CoDiCast is trained on a decade of ERA5 reanalysis data from the European Centre for Medium-Range Weather Forecasts (ECMWF). Experimental results demonstrate that our approach outperforms several existing data-driven methods in accuracy. Our conditional diffusion model, CoDiCast, can generate 3-day global weather forecasts, at 6-hour steps and $5.625^\circ$ latitude-longitude resolution, for over 5 variables, in about 12 minutes on a commodity A100 GPU machine with 80GB memory. The open-souced code is provided at \url{https://github.com/JimengShi/CoDiCast}.

Synthesizing accurate geometry and photo-realistic appearance of small scenes is an active area of research with compelling use cases in gaming, virtual reality, robotic-manipulation, autonomous driving, convenient product capture, and consumer-level photography. When applying scene geometry and appearance estimation techniques to robotics, we found that the narrow cone of possible viewpoints due to the limited range of robot motion and scene clutter caused current estimation techniques to produce poor quality estimates or even fail. On the other hand, in robotic applications, dense metric depth can often be measured directly using stereo and illumination can be controlled. Depth can provide a good initial estimate of the object geometry to improve reconstruction, while multi-illumination images can facilitate relighting. In this work we demonstrate a method to incorporate dense metric depth into the training of neural 3D representations and address an artifact observed while jointly refining geometry and appearance by disambiguating between texture and geometry edges. We also discuss a multi-flash stereo camera system developed to capture the necessary data for our pipeline and show results on relighting and view synthesis with a few training views., Comment: Project webpage: https://stereomfc.github.io

Large Language Models (LLMs) excel at generating human-like dialogues and comprehending text. However, understanding the subtleties of complex exchanges in language remains a challenge. We propose a bootstrapping framework that leverages self-generated feedback to enhance LLM reasoning capabilities for lie detection. The framework consists of three stages: suggestion, feedback collection, and modification. In the suggestion stage, a cost-effective language model generates initial predictions based on game state and dialogue. The feedback-collection stage involves a language model providing feedback on these predictions. In the modification stage, a more advanced language model refines the initial predictions using the auto-generated feedback. We investigate the application of the proposed framework for detecting betrayal and deception in Diplomacy games, and compare it with feedback from professional human players. The LLM-generated feedback exhibits superior quality and significantly enhances the performance of the model. Our approach achieves a 39% improvement over the zero-shot baseline in lying-F1 without the need for any training data, rivaling state-of-the-art supervised learning results., Comment: 19 pages, 18 figures

Collagenolytic degradation is a process fundamental to tissue remodeling. The microarchitecture of collagen fibril networks changes during development, aging, and disease. Such changes to microarchitecture are often accompanied by changes in matrix degradability. In vitro, collagen matrices of the same concentration but different microarchitectures also vary in degradation rate. How do different microarchitectures affect matrix degradation? To answer this question, we developed a computational model of collagen degradation. We first developed a lattice model that describes collagen degradation at the scale of a single fibril. We then extended this model to investigate the role of microarchitecture using Brownian dynamics simulation of enzymes in a multi-fibril three dimensional matrix to predict its degradability. Our simulations predict that the distribution of enzymes around the fibrils is non-uniform and depends on the microarchitecture of the matrix. This non-uniformity in enzyme distribution can lead to different extents of degradability for matrices of different microarchitectures. Our model predictions were tested using in vitro experiments with synthesized collagen gels of different microarchitectures. Experiments showed that indeed degradation of collagen depends on the matrix architecture and fibril thickness. In summary, our study shows that the microarchitecture of the collagen matrix is an important determinant of its degradability.

We present ShieldGemma, a comprehensive suite of LLM-based safety content moderation models built upon Gemma2. These models provide robust, state-of-the-art predictions of safety risks across key harm types (sexually explicit, dangerous content, harassment, hate speech) in both user input and LLM-generated output. By evaluating on both public and internal benchmarks, we demonstrate superior performance compared to existing models, such as Llama Guard (+10.8\% AU-PRC on public benchmarks) and WildCard (+4.3\%). Additionally, we present a novel LLM-based data curation pipeline, adaptable to a variety of safety-related tasks and beyond. We have shown strong generalization performance for model trained mainly on synthetic data. By releasing ShieldGemma, we provide a valuable resource to the research community, advancing LLM safety and enabling the creation of more effective content moderation solutions for developers.

Persona agents, which are LLM agents that act according to an assigned persona, have demonstrated impressive contextual response capabilities across various applications. These persona agents offer significant enhancements across diverse sectors, such as education, healthcare, and entertainment, where model developers can align agent responses to different user requirements thereby broadening the scope of agent applications. However, evaluating persona agent performance is incredibly challenging due to the complexity of assessing persona adherence in free-form interactions across various environments that are relevant to each persona agent. We introduce PersonaGym, the first dynamic evaluation framework for assessing persona agents, and PersonaScore, the first automated human-aligned metric grounded in decision theory for comprehensive large-scale evaluation of persona agents. Our evaluation of 6 open and closed-source LLMs, using a benchmark encompassing 200 personas and 10,000 questions, reveals significant opportunities for advancement in persona agent capabilities across state-of-the-art models. For example, Claude 3.5 Sonnet only has a 2.97% relative improvement in PersonaScore than GPT 3.5 despite being a much more advanced model. Importantly, we find that increased model size and complexity do not necessarily imply enhanced persona agent capabilities thereby highlighting the pressing need for algorithmic and architectural invention towards faithful and performant persona agents., Comment: 21 pages, 5 figures

Existing benchmarks do not test language agents on their interaction with human users or ability to follow domain-specific rules, both of which are vital for deploying them in real world applications. We propose $\tau$-bench, a benchmark emulating dynamic conversations between a user (simulated by language models) and a language agent provided with domain-specific API tools and policy guidelines. We employ an efficient and faithful evaluation process that compares the database state at the end of a conversation with the annotated goal state. We also propose a new metric (pass^k) to evaluate the reliability of agent behavior over multiple trials. Our experiments show that even state-of-the-art function calling agents (like gpt-4o) succeed on <50% of the tasks, and are quite inconsistent (pass^8 <25% in retail). Our findings point to the need for methods that can improve the ability of agents to act consistently and follow rules reliably.

Large Language Models trained on code corpora (code-LLMs) have demonstrated impressive performance in various coding assistance tasks. However, despite their increased size and training dataset, code-LLMs still have limitations such as suggesting codes with syntactic errors, variable misuse etc. Some studies argue that code-LLMs perform well on coding tasks because they use self-attention and hidden representations to encode relations among input tokens. However, previous works have not studied what code properties are not encoded by code-LLMs. In this paper, we conduct a fine-grained analysis of attention maps and hidden representations of code-LLMs. Our study indicates that code-LLMs only encode relations among specific subsets of input tokens. Specifically, by categorizing input tokens into syntactic tokens and identifiers, we found that models encode relations among syntactic tokens and among identifiers, but they fail to encode relations between syntactic tokens and identifiers. We also found that fine-tuned models encode these relations poorly compared to their pre-trained counterparts. Additionally, larger models with billions of parameters encode significantly less information about code than models with only a few hundred million parameters.

Perceiving and navigating through work zones is challenging and under-explored, even with major strides in self-driving research. An important reason is the lack of open datasets for developing new algorithms to address this long-tailed scenario. We propose the ROADWork dataset to learn how to recognize, observe and analyze and drive through work zones. We find that state-of-the-art foundation models perform poorly on work zones. With our dataset, we improve upon detecting work zone objects (+26.2 AP), while discovering work zones with higher precision (+32.5%) at a much higher discovery rate (12.8 times), significantly improve detecting (+23.9 AP) and reading (+14.2% 1-NED) work zone signs and describing work zones (+36.7 SPICE). We also compute drivable paths from work zone navigation videos and show that it is possible to predict navigational goals and pathways such that 53.6% goals have angular error (AE) < 0.5 degrees (+9.9 %) and 75.3% pathways have AE < 0.5 degrees (+8.1 %).

Reducing serving cost and latency is a fundamental concern for the deployment of language models (LMs) in business applications. To address this, cascades of LMs offer an effective solution that conditionally employ smaller models for simpler queries. Cascaded systems are typically built with independently trained models, neglecting the advantages of considering inference-time interactions of the cascaded LMs during training. In this paper, we present cascade-aware training(CAT), an approach to optimizing the overall quality-cost performance tradeoff of a cascade of LMs. We achieve inference-time benefits by training the small LM with awareness of its place in a cascade and downstream capabilities. We demonstrate the value of the proposed method with over 60 LM tasks of the SuperGLUE, WMT22, and FLAN2021 datasets., Comment: 22 pages, 13 figures

Cascades and speculative decoding are two common approaches to improving language models' inference efficiency. Both approaches involve interleaving models of different sizes, but via fundamentally distinct mechanisms: cascades employ a deferral rule that invokes the larger model only for "hard" inputs, while speculative decoding uses speculative execution to primarily invoke the larger model in parallel verification mode. These mechanisms offer different benefits: empirically, cascades offer better cost-quality trade-offs, often even outperforming the large model, while theoretically, speculative decoding offers a guarantee of quality-neutrality. In this paper, we leverage the best of both these approaches by designing new speculative cascading techniques that implement their deferral rule through speculative execution. We characterize the optimal deferral rule for our speculative cascades, and employ a plug-in approximation to the optimal rule. Experiments with Gemma and T5 models on a range of language benchmarks show that our approach yields better cost quality trade-offs than cascading and speculative decoding baselines.

Language model (LM) agents are increasingly being used to automate complicated tasks in digital environments. Just as humans benefit from powerful software applications, such as integrated development environments, for complex tasks like software engineering, we posit that LM agents represent a new category of end users with their own needs and abilities, and would benefit from specially-built interfaces to the software they use. We investigate how interface design affects the performance of language model agents. As a result of this exploration, we introduce SWE-agent: a system that facilitates LM agents to autonomously use computers to solve software engineering tasks. SWE-agent's custom agent-computer interface (ACI) significantly enhances an agent's ability to create and edit code files, navigate entire repositories, and execute tests and other programs. We evaluate SWE-agent on SWE-bench and HumanEvalFix, achieving state-of-the-art performance on both with a pass@1 rate of 12.5% and 87.7%, respectively, far exceeding the previous state-of-the-art achieved with non-interactive LMs. Finally, we provide insight on how the design of the ACI can impact agents' behavior and performance., Comment: Code, data, and demo available at https://swe-agent.com

We introduce Constellation, a dataset of 13K images suitable for research on detection of objects in dense urban streetscapes observed from high-elevation cameras, collected for a variety of temporal conditions. The dataset addresses the need for curated data to explore problems in small object detection exemplified by the limited pixel footprint of pedestrians observed tens of meters from above. It enables the testing of object detection models for variations in lighting, building shadows, weather, and scene dynamics. We evaluate contemporary object detection architectures on the dataset, observing that state-of-the-art methods have lower performance in detecting small pedestrians compared to vehicles, corresponding to a 10% difference in average precision (AP). Using structurally similar datasets for pretraining the models results in an increase of 1.8% mean AP (mAP). We further find that incorporating domain-specific data augmentations helps improve model performance. Using pseudo-labeled data, obtained from inference outcomes of the best-performing models, improves the performance of the models. Finally, comparing the models trained using the data collected in two different time intervals, we find a performance drift in models due to the changes in intersection conditions over time. The best-performing model achieves a pedestrian AP of 92.0% with 11.5 ms inference time on NVIDIA A100 GPUs, and an mAP of 95.4%.

Computing olympiads contain some of the most challenging problems for humans, requiring complex algorithmic reasoning, puzzle solving, in addition to generating efficient code. However, it has been understudied as a domain to evaluate language models (LMs). In this paper, we introduce the USACO benchmark with 307 problems from the USA Computing Olympiad, along with high-quality unit tests, reference code, and official analyses for each problem. These resources enable us to construct and test a range of LM inference methods for competitive programming for the first time. We find GPT-4 only achieves a 8.7% pass@1 accuracy with zero-shot chain-of-thought prompting, and our best inference method improves it to 20.2% using a combination of self-reflection and retrieval over episodic knowledge. However, this is far from solving the benchmark. To better understand the remaining challenges, we design a novel human-in-the-loop study and surprisingly find that a small number of targeted hints enable GPT-4 to solve 13 out of 15 problems previously unsolvable by any model and method. Our benchmark, baseline methods, quantitative results, and qualitative analysis serve as an initial step toward LMs with grounded, creative, and algorithmic reasoning., Comment: Code and data: https://princeton-nlp.github.io/USACOBench/

Recent advances in language models (LMs) have led to significant improvements in quality on complex NLP tasks, but at the expense of increased inference costs. Cascading offers a simple strategy to achieve more favorable cost-quality tradeoffs: here, a small model is invoked for most "easy" instances, while a few "hard" instances are deferred to the large model. While the principles underpinning cascading are well-studied for classification tasks - with deferral based on predicted class uncertainty favored theoretically and practically - a similar understanding is lacking for generative LM tasks. In this work, we initiate a systematic study of deferral rules for LM cascades. We begin by examining the natural extension of predicted class uncertainty to generative LM tasks, namely, the predicted sequence uncertainty. We show that this measure suffers from the length bias problem, either over- or under-emphasizing outputs based on their lengths. This is because LMs produce a sequence of uncertainty values, one for each output token; and moreover, the number of output tokens is variable across examples. To mitigate this issue, we propose to exploit the richer token-level uncertainty information implicit in generative LMs. We argue that naive predicted sequence uncertainty corresponds to a simple aggregation of these uncertainties. By contrast, we show that incorporating token-level uncertainty through learned post-hoc deferral rules can significantly outperform such simple aggregation strategies, via experiments on a range of natural language benchmarks with FLAN-T5 models. We further show that incorporating embeddings from the smaller model and intermediate layers of the larger model can give an additional boost in the overall cost-quality tradeoff.

State-of-the-art large language models (LLMs) have become indispensable tools for various tasks. However, training LLMs to serve as effective assistants for humans requires careful consideration. A promising approach is reinforcement learning from human feedback (RLHF), which leverages human feedback to update the model in accordance with human preferences and mitigate issues like toxicity and hallucinations. Yet, an understanding of RLHF for LLMs is largely entangled with initial design choices that popularized the method and current research focuses on augmenting those choices rather than fundamentally improving the framework. In this paper, we analyze RLHF through the lens of reinforcement learning principles to develop an understanding of its fundamentals, dedicating substantial focus to the core component of RLHF -- the reward model. Our study investigates modeling choices, caveats of function approximation, and their implications on RLHF training algorithms, highlighting the underlying assumptions made about the expressivity of reward. Our analysis improves the understanding of the role of reward models and methods for their training, concurrently revealing limitations of the current methodology. We characterize these limitations, including incorrect generalization, model misspecification, and the sparsity of feedback, along with their impact on the performance of a language model. The discussion and analysis are substantiated by a categorical review of current literature, serving as a reference for researchers and practitioners to understand the challenges of RLHF and build upon existing efforts.

As factories continue to evolve into collaborative spaces with multiple robots working together with human supervisors in the loop, ensuring safety for all actors involved becomes critical. Currently, laser-based light curtain sensors are widely used in factories for safety monitoring. While these conventional safety sensors meet high accuracy standards, they are difficult to reconfigure and can only monitor a fixed user-defined region of space. Furthermore, they are typically expensive. Instead, we leverage a controllable depth sensor, programmable light curtains (PLC), to develop an inexpensive and flexible real-time safety monitoring system for collaborative robot workspaces. Our system projects virtual dynamic safety envelopes that tightly envelop the moving robot at all times and detect any objects that intrude the envelope. Furthermore, we develop an instrumentation algorithm that optimally places (multiple) PLCs in a workspace to maximize the visibility coverage of robots. Our work enables fence-less human-robot collaboration, while scaling to monitor multiple robots with few sensors. We analyze our system in a real manufacturing testbed with four robot arms and demonstrate its capabilities as a fast, accurate, and inexpensive safety monitoring solution., Comment: Under review for IROS '24. Webpage http://cmu-mfi.github.io/plc-safety

Current methods for 2D and 3D object understanding struggle with severe occlusions in busy urban environments, partly due to the lack of large-scale labeled ground-truth annotations for learning occlusion. In this work, we introduce a novel framework for automatically generating a large, realistic dataset of dynamic objects under occlusions using freely available time-lapse imagery. By leveraging off-the-shelf 2D (bounding box, segmentation, keypoint) and 3D (pose, shape) predictions as pseudo-groundtruth, unoccluded 3D objects are identified automatically and composited into the background in a clip-art style, ensuring realistic appearances and physically accurate occlusion configurations. The resulting clip-art image with pseudo-groundtruth enables efficient training of object reconstruction methods that are robust to occlusions. Our method demonstrates significant improvements in both 2D and 3D reconstruction, particularly in scenarios with heavily occluded objects like vehicles and people in urban scenes., Comment: To appear in CVPR 2024. Homepage: https://www.cs.cmu.edu/~walt3d

Driving is challenging in conditions like night, rain, and snow. Lack of good labeled datasets has hampered progress in scene understanding under such conditions. Unsupervised Domain Adaptation (UDA) using large labeled clear-day datasets is a promising research direction in such cases. However, many UDA methods are trained with dominant scene backgrounds (e.g., roads, sky, sidewalks) that appear dramatically different across domains. As a result, they struggle to learn effective features of smaller and often sparse foreground objects (e.g., people, vehicles, signs). In this work, we improve UDA training by applying in-place image warping to focus on salient objects. We design instance-level saliency guidance to adaptively oversample object regions and undersample background areas, which reduces adverse effects from background context and enhances backbone feature learning. Our approach improves adaptation across geographies, lighting, and weather conditions, and is agnostic to the task (segmentation, detection), domain adaptation algorithm, saliency guidance, and underlying model architecture. Result highlights include +6.1 mAP50 for BDD100K Clear $\rightarrow$ DENSE Foggy, +3.7 mAP50 for BDD100K Day $\rightarrow$ Night, +3.0 mAP50 for BDD100K Clear $\rightarrow$ Rainy, and +6.3 mIoU for Cityscapes $\rightarrow$ ACDC. Besides, Our method adds minimal training memory and no additional inference latency. Code is available at https://github.com/ShenZheng2000/Instance-Warp, Comment: WACV 2025 Accepted Paper

In this work, we address two limitations of existing conditional diffusion models: their slow inference speed due to the iterative denoising process and their reliance on paired data for model fine-tuning. To tackle these issues, we introduce a general method for adapting a single-step diffusion model to new tasks and domains through adversarial learning objectives. Specifically, we consolidate various modules of the vanilla latent diffusion model into a single end-to-end generator network with small trainable weights, enhancing its ability to preserve the input image structure while reducing overfitting. We demonstrate that, for unpaired settings, our model CycleGAN-Turbo outperforms existing GAN-based and diffusion-based methods for various scene translation tasks, such as day-to-night conversion and adding/removing weather effects like fog, snow, and rain. We extend our method to paired settings, where our model pix2pix-Turbo is on par with recent works like Control-Net for Sketch2Photo and Edge2Image, but with a single-step inference. This work suggests that single-step diffusion models can serve as strong backbones for a range of GAN learning objectives. Our code and models are available at https://github.com/GaParmar/img2img-turbo., Comment: Github: https://github.com/GaParmar/img2img-turbo

Routine inspections for critical infrastructures such as bridges are required in most jurisdictions worldwide. Such routine inspections are largely visual in nature, which are qualitative, subjective, and not repeatable. Although robotic infrastructure inspections address such limitations, they cannot replace the superior ability of experts to make decisions in complex situations, thus making human-robot interaction systems a promising technology. This study presents a novel gaze-based human-robot interaction system, designed to augment the visual inspection performance through mixed reality. Through holograms from a mixed reality device, gaze can be utilized effectively to estimate the properties of the defect in real-time. Additionally, inspectors can monitor the inspection progress online, which enhances the speed of the entire inspection process. Limited controlled experiments demonstrate its effectiveness across various users and defect types. To our knowledge, this is the first demonstration of the real-time application of eye gaze in civil infrastructure inspections., Comment: 7 pages, 8 figures, 1 supplementary video; Accepted to the 2024 IEEE International Conference on Robotics and Automation (ICRA)

Large language models (LLMs) have shown strong results on a range of applications, including regression and scoring tasks. Typically, one obtains outputs from an LLM via autoregressive sampling from the model's output distribution. We show that this inference strategy can be sub-optimal for common regression and scoring evaluation metrics. As a remedy, we build on prior work on Minimum Bayes Risk decoding, and propose alternate inference strategies that estimate the Bayes-optimal solution for regression and scoring metrics in closed-form from sampled responses. We show that our proposal significantly improves over baselines across datasets and models., Comment: EMNLP Findings 2024