Large Language Models (LLMs) demonstrate promising capabilities in solving simple scientific problems but often produce hallucinations for complex ones. While integrating LLMs with tools can increase reliability, this approach typically results in over-reliance on tools, diminishing the model's ability to solve simple problems through basic reasoning. In contrast, human experts first assess problem complexity using domain knowledge before choosing an appropriate solution approach. Inspired by this human problem-solving process, we propose a novel two-component fine-tuning method. In the first component World Knowledge Distillation (WKD), LLMs learn directly from solutions generated using tool's information to internalize domain knowledge. In the second component Tool Usage Adaptation (TUA), we partition problems into easy and hard categories based on the model's direct answering accuracy. While maintaining the same alignment target for easy problems as in WKD, we train the model to intelligently switch to tool usage for more challenging problems. We validate our method on six scientific benchmark datasets, spanning mathematics, climate science and epidemiology. On average, our models demonstrate a 28.18% improvement in answer accuracy and a 13.89% increase in tool usage precision across all datasets, surpassing state-of-the-art models including GPT-4o and Claude-3.5.

In-context learning (ICL) is an effective approach to help large language models (LLMs) adapt to various tasks by providing demonstrations of the target task. Considering the high cost of labeling demonstrations, many methods propose synthesizing demonstrations from scratch using LLMs. However, the quality of the demonstrations synthesized from scratch is limited by the capabilities and knowledge of LLMs. To address this, inspired by transfer learning, we propose In-Context Transfer Learning (ICTL), which synthesizes target task demonstrations by transferring labeled demonstrations from similar source tasks. ICTL consists of two steps: source sampling and target transfer. First, we define an optimization objective, which minimizes transfer error to sample source demonstrations similar to the target task. Then, we employ LLMs to transfer the sampled source demonstrations to the target task, matching the definition and format of the target task. Experiments on Super-NI show that ICTL outperforms synthesis from scratch by 2.0% on average, demonstrating the effectiveness of our method In-context learning (ICL) is an effective approach for large language models (LLMs) to adapt to various tasks based on the brilliant generalize ability of LLMs (Xun et al., 2017; Song et al., 2023b; Luo et al., 2024a). During the inference with ICL, input not only includes user questions but also several demonstrations to guide LLMs in generating answers correctly. Considering the high cost of demonstration labeling, many methods utilize LLMs to synthesize demonstrations from scratch without human involvement (Kim et al., 2022; Jin & Lu, 2024). For instance, Self-ICL (Chen et al., 2023b) employs LLMs to synthesize demonstration based on the task definition, while Su et al. (2024) improves the synthesis through iterations, where each iteration uses the previous results. However, the synthesis using LLMs from scratch is constrained by the capabilities and knowledge of LLMs, limiting the quality of the synthesized demonstrations (Yu et al., 2023).

However, Moreover, the failure to acknowledge the socio-cultural concerns persist regarding the transparency and context context within which data is produced can introduce biases of data collection methodologies, especially when sourced into datasets. For example, algorithms trained on datasets through crowdsourcing platforms. Crowdsourcing often devoid of the historical context of segregation may inadvertently employs low-wage workers with poor working conditions perpetuate biases against certain minority groups and lacks consideration for the representativeness of annotators, [12]. Furthermore, the identities of workers involved in annotations leading to algorithms that fail to represent diverse are frequently overlooked, leading to a lack of diversity views and perpetuate biases against certain groups. To address in viewpoints captured within datasets. This bias is these limitations, we propose a methodology involving compounded by the common practice of aggregating annotations a co-design model that actively engages stakeholders at key through majority voting [5].

Retrieval-Augmented Generation (RAG) has emerged as a reliable external knowledge augmentation technique to mitigate hallucination issues and parameterized knowledge limitations in Large Language Models (LLMs). Existing Adaptive RAG (ARAG) systems struggle to effectively explore multiple retrieval sources due to their inability to select the right source at the right time. To address this, we propose a multi-source ARAG framework, termed MSPR, which synergizes reasoning and preference-driven retrieval to adaptive decide "when and what to retrieve" and "which retrieval source to use". To better adapt to retrieval sources of differing characteristics, we also employ retrieval action adjustment and answer feedback strategy. They enable our framework to fully explore the high-quality primary source while supplementing it with secondary sources at the right time. Extensive and multi-dimensional experiments conducted on three datasets demonstrate the superiority and effectiveness of MSPR.

Self-improvement methods enable large language models (LLMs) to generate solutions themselves and iteratively train on filtered, high-quality rationales. This process proves effective and reduces the reliance on human supervision in LLMs' reasoning, but the performance soon plateaus. We delve into the process and find that models tend to over-sample on easy queries and under-sample on queries they have yet to master. As iterations proceed, this imbalance in sampling is exacerbated, leading to a long-tail distribution where solutions to difficult queries almost diminish. This phenomenon limits the performance gain of self-improving models. A straightforward solution is brute-force sampling to balance the distribution, which significantly raises computational costs. In this paper, we introduce Guided Self-Improvement (GSI), a strategy aimed at improving the efficiency of sampling challenging heavy-tailed data. It leverages Socratic-style guidance signals to help LLM reasoning with complex queries, reducing the exploration effort and minimizing computational overhead. Experiments on four models across diverse mathematical tasks show that GSI strikes a balance between performance and efficiency, while also being effective on held-out tasks.

This work introduces a novel paradigm for solving optimal control problems for hybrid dynamical systems under uncertainties. Robotic systems having contact with the environment can be modeled as hybrid systems. Controller design for hybrid systems under disturbances is complicated by the discontinuous jump dynamics, mode changes with inconsistent state dimensions, and variations in jumping timing and states caused by noise. We formulate this problem into a stochastic control problem with hybrid transition constraints and propose the Hybrid Path Integral (H-PI) framework to obtain the optimal controller. Despite random mode changes across stochastic path samples, we show that the ratio between hybrid path distributions with varying drift terms remains analogous to the smooth path distributions. We then show that the optimal controller can be obtained by evaluating a path integral with hybrid constraints. Importance sampling for path distributions with hybrid dynamics constraints is introduced to reduce the variance of the path integral evaluation, where we leverage the recently developed Hybrid iterative-Linear-Quadratic-Regulator (H-iLQR) controller to induce a hybrid path distribution proposal with low variance. The proposed method is validated through numerical experiments on various hybrid systems and extensive ablation studies. All the sampling processes are conducted in parallel on a Graphics Processing Unit (GPU).

Narratives are key interpretative devices by which humans make sense of political reality. As the significance of narratives for understanding current societal issues such as polarization and misinformation becomes increasingly evident, there is a growing demand for methods that support their empirical analysis. To this end, we propose a graph-based formalism and machine-guided method for extracting, representing, and analyzing selected narrative signals from digital textual corpora, based on Abstract Meaning Representation (AMR). The formalism and method introduced here specifically cater to the study of political narratives that figure in texts from digital media such as archived political speeches, social media posts, political manifestos and transcripts of parliamentary debates. We conceptualize these political narratives as a type of ontological narratives: stories by which actors position themselves as political beings, and which are akin to political worldviews in which actors present their normative vision of the world, or aspects thereof. We approach the study of such political narratives as a problem of information retrieval: starting from a textual corpus, we first extract a graph-like representation of the meaning of each sentence in the corpus using AMR. Drawing on transferable concepts from narratology, we then apply a set of heuristics to filter these graphs for representations of 1) actors, 2) the events in which these actors figure, and 3) traces of the perspectivization of these events. We approach these references to actors, events, and instances of perspectivization as core narrative signals that initiate a further analysis by alluding to larger political narratives. By means of a case study of State of the European Union addresses, we demonstrate how the formalism can be used to inductively surface signals of political narratives from public discourse.

We present TaxaBind, a unified embedding space for characterizing any species of interest. TaxaBind is a multimodal embedding space across six modalities: ground-level images of species, geographic location, satellite image, text, audio, and environmental features, useful for solving ecological problems. To learn this joint embedding space, we leverage ground-level images of species as a binding modality. We propose multimodal patching, a technique for effectively distilling the knowledge from various modalities into the binding modality. We construct two large datasets for pretraining: iSatNat with species images and satellite images, and iSoundNat with species images and audio. Additionally, we introduce TaxaBench-8k, a diverse multimodal dataset with six paired modalities for evaluating deep learning models on ecological tasks. Experiments with TaxaBind demonstrate its strong zero-shot and emergent capabilities on a range of tasks including species classification, cross-model retrieval, and audio classification. The datasets and models are made available at https://github.com/mvrl/TaxaBind.

Platooning or cooperative adaptive cruise control (CACC) has been investigated for decades, but debate about its lasting impact is still ongoing. Even though platooning benefits and platoon formation are rather well understood for trucks, this is less clear for passenger cars, which have a higher heterogeneity in trips and drivers' preferences. Most importantly, it remains unclear how to form platoons of passenger cars in order to optimize the personal benefit for the individual driver. To this end, in this paper, we propose a novel platoon formation algorithm that optimizes the personal benefit for drivers of individual passenger cars. For computing vehicle-to-platoon assignments, the algorithm utilizes a new metric that we propose to evaluate the personal benefits of various driving systems, including platooning. By combining fuel and travel time costs into a single monetary value, drivers can estimate overall trip costs according to a personal monetary value for time spent. This provides an intuitive way for drivers to understand and compare the benefits of driving systems like human driving, adaptive cruise control (ACC), and, of course, platooning. Unlike previous similarity-based methods, our proposed algorithm forms platoons only when beneficial for the driver, rather than for the sake of platooning only. Results of a large-scale simulation study demonstrate that our proposed algorithm outperforms normal ACC as well as previous similarity-based platooning approaches by balancing fuel savings and travel time, independent of traffic and drivers' time cost.

Graphs model complex relationships between entities, with nodes and edges capturing intricate connections. Node representation learning involves transforming nodes into low-dimensional embeddings. These embeddings are typically used as features for downstream tasks. Therefore, their quality has a significant impact on task performance. Existing approaches for node representation learning span (semi-)supervised, unsupervised, and self-supervised paradigms. In graph domains, (semi-)supervised learning often only optimizes models based on class labels, neglecting other abundant graph signals, which limits generalization. While self-supervised or unsupervised learning produces representations that better capture underlying graph signals, the usefulness of these captured signals for downstream target tasks can vary. To bridge this gap, we introduce Target-Aware Contrastive Learning (Target-aware CL) which aims to enhance target task performance by maximizing the mutual information between the target task and node representations with a self-supervised learning process. This is achieved through a sampling function, XGBoost Sampler (XGSampler), to sample proper positive examples for the proposed Target-Aware Contrastive Loss (XTCL). By minimizing XTCL, Target-aware CL increases the mutual information between the target task and node representations, such that model generalization is improved. Additionally, XGSampler enhances the interpretability of each signal by showing the weights for sampling the proper positive examples. We show experimentally that XTCL significantly improves the performance on two target tasks: node classification and link prediction tasks, compared to state-of-the-art models.