Search

Effective field theories are constrained by the requirement that their constituents never move superluminally on non-trivial backgrounds. In this paper, we study time delays experienced by photons propagating on charged shockwave backgrounds in five dimensions. In the absence of gravity -- where the shockwaves are electric fields sourced by boosted charges -- we derive positivity bounds for the four-derivative corrections to electromagnetism, reproducing previous results derived from scattering amplitudes. By considering the gravitational shockwaves sourced by Reissner-Nordstr\"om black holes, we derive new constraints in the presence of gravity. We observe the by-now familiar weakening of positivity bounds in the presence of gravity, but without the logarithmic divergences present in 4d. We find that the strongest bounds appear by examining the time delay near the horizon of the smallest possible black hole, and discuss on the validity of the EFT expansion in this region. We comment on our bounds in the context of the swampland program as well as their relation with the positivity bounds obtained from dispersion relations., Comment: 31 pages, 1 figure

Unlabeled motion planning involves assigning a set of robots to target locations while ensuring collision avoidance, aiming to minimize the total distance traveled. The problem forms an essential building block for multi-robot systems in applications such as exploration, surveillance, and transportation. We address this problem in a decentralized setting where each robot knows only the positions of its $k$-nearest robots and $k$-nearest targets. This scenario combines elements of combinatorial assignment and continuous-space motion planning, posing significant scalability challenges for traditional centralized approaches. To overcome these challenges, we propose a decentralized policy learned via a Graph Neural Network (GNN). The GNN enables robots to determine (1) what information to communicate to neighbors and (2) how to integrate received information with local observations for decision-making. We train the GNN using imitation learning with the centralized Hungarian algorithm as the expert policy, and further fine-tune it using reinforcement learning to avoid collisions and enhance performance. Extensive empirical evaluations demonstrate the scalability and effectiveness of our approach. The GNN policy trained on 100 robots generalizes to scenarios with up to 500 robots, outperforming state-of-the-art solutions by 8.6\% on average and significantly surpassing greedy decentralized methods. This work lays the foundation for solving multi-robot coordination problems in settings where scalability is important., Comment: 6 pages, 6 figures, submitted to ICRA 2025

We study moduli stabilization via fluxes in the $2^6$ Landau-Ginzburg model. Fluxes not only give masses to scalar fields but can also induce higher order couplings that stabilize massless fields. We investigate this for several different flux choices in the $2^6$ model and find two examples that are inconsistent with the Refined Tadpole Conjecture. We also present, to our knowledge, the first 4d $\mathcal{N}=1$ Minkowski solution in string theory without any flat direction., Comment: 33 pages, 1 figure

Enterprise chatbots, powered by generative AI, are emerging as key applications to enhance employee productivity. Retrieval Augmented Generation (RAG), Large Language Models (LLMs), and orchestration frameworks like Langchain and Llamaindex are crucial for building these chatbots. However, creating effective enterprise chatbots is challenging and requires meticulous RAG pipeline engineering. This includes fine-tuning embeddings and LLMs, extracting documents from vector databases, rephrasing queries, reranking results, designing prompts, honoring document access controls, providing concise responses, including references, safeguarding personal information, and building orchestration agents. We present a framework for building RAG-based chatbots based on our experience with three NVIDIA chatbots: for IT/HR benefits, financial earnings, and general content. Our contributions are three-fold: introducing the FACTS framework (Freshness, Architectures, Cost, Testing, Security), presenting fifteen RAG pipeline control points, and providing empirical results on accuracy-latency tradeoffs between large and small LLMs. To the best of our knowledge, this is the first paper of its kind that provides a holistic view of the factors as well as solutions for building secure enterprise-grade chatbots.", Comment: 8 pages, 6 figures, 2 tables, Preprint submission to ACM CIKM 2024

Opinion summarization in e-commerce encapsulates the collective views of numerous users about a product based on their reviews. Typically, a product on an e-commerce platform has thousands of reviews, each review comprising around 10-15 words. While Large Language Models (LLMs) have shown proficiency in summarization tasks, they struggle to handle such a large volume of reviews due to context limitations. To mitigate, we propose a scalable framework called Xl-OpSumm that generates summaries incrementally. However, the existing test set, AMASUM has only 560 reviews per product on average. Due to the lack of a test set with thousands of reviews, we created a new test set called Xl-Flipkart by gathering data from the Flipkart website and generating summaries using GPT-4. Through various automatic evaluations and extensive analysis, we evaluated the framework's efficiency on two datasets, AMASUM and Xl-Flipkart. Experimental results show that our framework, Xl-OpSumm powered by Llama-3-8B-8k, achieves an average ROUGE-1 F1 gain of 4.38% and a ROUGE-L F1 gain of 3.70% over the next best-performing model.

Recent work on flux compactifications suggests that the tadpole constraint generically allows only a limited number of complex structure moduli to become massive, i.e., be stabilized at quadratic order in the spacetime superpotential. We study the effects of higher-order terms systematically around the Fermat point in the $1^9$ Landau-Ginzburg model. This model lives at strong coupling and features no K\"ahler moduli. We show that, depending on the flux, several massless fields can indeed be stabilized in this fashion, and argue that this paves the way to explicit ${\mathcal N}=1$ Minkowski vacua without flat directions. Along the way, we complete the classification of integral flux vectors with small tadpole contribution. Thereby we are closing in on a future complete understanding of all possible flux configurations in the $1^9$ Landau-Ginzburg model., Comment: 41 pages, 2 figures

In e-commerce, opinion summarization is the process of summarizing the consensus opinions found in product reviews. However, the potential of additional sources such as product description and question-answers (QA) has been considered less often. Moreover, the absence of any supervised training data makes this task challenging. To address this, we propose a novel synthetic dataset creation (SDC) strategy that leverages information from reviews as well as additional sources for selecting one of the reviews as a pseudo-summary to enable supervised training. Our Multi-Encoder Decoder framework for Opinion Summarization (MEDOS) employs a separate encoder for each source, enabling effective selection of information while generating the summary. For evaluation, due to the unavailability of test sets with additional sources, we extend the Amazon, Oposum+, and Flipkart test sets and leverage ChatGPT to annotate summaries. Experiments across nine test sets demonstrate that the combination of our SDC approach and MEDOS model achieves on average a 14.5% improvement in ROUGE-1 F1 over the SOTA. Moreover, comparative analysis underlines the significance of incorporating additional sources for generating more informative summaries. Human evaluations further indicate that MEDOS scores relatively higher in coherence and fluency with 0.41 and 0.5 (-1 to 1) respectively, compared to existing models. To the best of our knowledge, we are the first to generate opinion summaries leveraging additional sources in a self-supervised setting.

There is a growing need for Large Language Models (LLMs) to effectively use tools and external Application Programming Interfaces (APIs) to plan and complete tasks. As such, there is tremendous interest in methods that can acquire sufficient quantities of train and test data that involve calls to tools / APIs. Two lines of research have emerged as the predominant strategies for addressing this challenge. The first has focused on synthetic data generation techniques, while the second has involved curating task-adjacent datasets which can be transformed into API / Tool-based tasks. In this paper, we focus on the task of identifying, curating, and transforming existing datasets and, in turn, introduce API-BLEND, a large corpora for training and systematic testing of tool-augmented LLMs. The datasets mimic real-world scenarios involving API-tasks such as API / tool detection, slot filling, and sequencing of the detected APIs. We demonstrate the utility of the API-BLEND dataset for both training and benchmarking purposes., Comment: Accepted at ACL'24-main conference

Reinforcement Learning from Human Feedback (RLHF) has become a dominating strategy in aligning Language Models (LMs) with human values/goals. The key to the strategy is learning a reward model ($\varphi$), which can reflect the latent reward model of humans. While this strategy has proven effective, the training methodology requires a lot of human preference annotation (usually in the order of tens of thousands) to train $\varphi$. Such a large-scale annotation is justifiable when it's a one-time effort, and the reward model is universally applicable. However, human goals are subjective and depend on the task, requiring task-specific preference annotations, which can be impractical to fulfill. To address this challenge, we propose a novel approach to infuse domain knowledge into $\varphi$, which reduces the amount of preference annotation required ($21\times$), omits Alignment Tax, and provides some interpretability. We validate our approach in E-Commerce Opinion Summarization, with a significant reduction in dataset size (to just $940$ samples) while advancing the SOTA ($\sim4$ point ROUGE-L improvement, $68\%$ of times preferred by humans over SOTA). Our contributions include a novel Reward Modeling technique and two new datasets: PromptOpinSumm (supervised data for Opinion Summarization) and OpinPref (a gold-standard human preference dataset). The proposed methodology opens up avenues for efficient RLHF, making it more adaptable to applications with varying human values. We release the artifacts (Code: github.com/efficient-rlhf. PromptOpinSumm: hf.co/prompt-opin-summ. OpinPref: hf.co/opin-pref) for usage under MIT License., Comment: 19 pages, 6 figures, 21 tables

Social commerce platforms are emerging businesses where producers sell products through re-sellers who advertise the products to other customers in their social network. Due to the increasing popularity of this business model, thousands of small producers and re-sellers are starting to depend on these platforms for their livelihood; thus, it is important to provide fair earning opportunities to them. The enormous product space in such platforms prohibits manual search, and motivates the need for recommendation algorithms to effectively allocate product exposure and, consequently, earning opportunities. In this work, we focus on the fairness of such allocations in social commerce platforms and formulate the problem of assigning products to re-sellers as a fair division problem with indivisible items under two-sided cardinality constraints, wherein each product must be given to at least a certain number of re-sellers and each re-seller must get a certain number of products. Our work systematically explores various well-studied benchmarks of fairness -- including Nash social welfare, envy-freeness up to one item (EF1), and equitability up to one item (EQ1) -- from both theoretical and experimental perspectives. We find that the existential and computational guarantees of these concepts known from the unconstrained setting do not extend to our constrained model. To address this limitation, we develop a mixed-integer linear program and other scalable heuristics that provide near-optimal approximation of Nash social welfare in simulated and real social commerce datasets. Overall, our work takes the first step towards achieving provable fairness alongside reasonable revenue guarantees on social commerce platforms.

Evaluation of opinion summaries using conventional reference-based metrics rarely provides a holistic evaluation and has been shown to have a relatively low correlation with human judgments. Recent studies suggest using Large Language Models (LLMs) as reference-free metrics for NLG evaluation, however, they remain unexplored for opinion summary evaluation. Moreover, limited opinion summary evaluation datasets inhibit progress. To address this, we release the SUMMEVAL-OP dataset covering 7 dimensions related to the evaluation of opinion summaries: fluency, coherence, relevance, faithfulness, aspect coverage, sentiment consistency, and specificity. We investigate Op-I-Prompt a dimension-independent prompt, and Op-Prompts, a dimension-dependent set of prompts for opinion summary evaluation. Experiments indicate that Op-I-Prompt emerges as a good alternative for evaluating opinion summaries achieving an average Spearman correlation of 0.70 with humans, outperforming all previous approaches. To the best of our knowledge, we are the first to investigate LLMs as evaluators on both closed-source and open-source models in the opinion summarization domain.