Showing 1–11 of 11 results for author: Lastras, L A

ToolRM: Outcome Reward Models for Tool-Calling Large Language Models

Authors: Mayank Agarwal, Ibrahim Abdelaziz, Kinjal Basu, Merve Unuvar, Luis A. Lastras, Yara Rizk, Pavan Kapanipathi

Abstract: As large language models (LLMs) increasingly interact with external tools, reward modeling for tool use has emerged as a critical yet underexplored area of research. Existing reward models, trained primarily on natural language outputs, struggle to evaluate tool-based reasoning and execution. To quantify this gap, we introduce FC-RewardBench, the first benchmark to systematically evaluate reward m… ▽ More As large language models (LLMs) increasingly interact with external tools, reward modeling for tool use has emerged as a critical yet underexplored area of research. Existing reward models, trained primarily on natural language outputs, struggle to evaluate tool-based reasoning and execution. To quantify this gap, we introduce FC-RewardBench, the first benchmark to systematically evaluate reward models in tool-calling scenarios. Our analysis shows that current reward models frequently miss key signals of effective tool use, highlighting the need for domain-specific modeling. We address this by proposing a training framework for outcome reward models using data synthesized from permissively licensed, open-weight LLMs. We introduce ToolRM - a suite of reward models for tool-use ranging from 1.7B to 14B parameters. Across diverse settings, these models consistently outperform general-purpose baselines. Notably, they achieve up to a 25% improvement with Best-of-N sampling, while also improving robustness to input noise, enabling effective data filtering, and supporting RL-training of policy models. △ Less

Submitted 7 January, 2026; v1 submitted 15 September, 2025; originally announced September 2025.

Breaking through the classical Shannon entropy limit: A new frontier through logical semantics

Authors: Luis A. Lastras, Barry M. Trager, Jonathan Lenchner, Wojciech Szpankowski, Chai Wah Wu, Mark S. Squillante, Alexander Gray

Abstract: Information theory has provided foundations for the theories of several application areas critical for modern society, including communications, computer storage, and AI. A key aspect of Shannon's 1948 theory is a sharp lower bound on the number of bits needed to encode and communicate a string of symbols. When he introduced the theory, Shannon famously excluded any notion of semantics behind the… ▽ More Information theory has provided foundations for the theories of several application areas critical for modern society, including communications, computer storage, and AI. A key aspect of Shannon's 1948 theory is a sharp lower bound on the number of bits needed to encode and communicate a string of symbols. When he introduced the theory, Shannon famously excluded any notion of semantics behind the symbols being communicated. This semantics-free notion went on to have massive impact on communication and computing technologies, even as multiple proposals for reintroducing semantics in a theory of information were being made, notably one where Carnap and Bar-Hillel used logic and reasoning to capture semantics. In this paper we present, for the first time, a Shannon-style analysis of a communication system equipped with a deductive reasoning capability, implemented using logical inference. We use some of the most important techniques developed in information theory to demonstrate significant and sometimes surprising gains in communication efficiency availed to us through such capability, demonstrated also through practical codes. We thus argue that proposals for a semantic information theory should include the power of deductive reasoning to magnify the value of transmitted bits as we strive to fully unlock the inherent potential of semantics. △ Less

Submitted 31 December, 2024; originally announced January 2025.

NESTFUL: A Benchmark for Evaluating LLMs on Nested Sequences of API Calls

Authors: Kinjal Basu, Ibrahim Abdelaziz, Kiran Kate, Mayank Agarwal, Maxwell Crouse, Yara Rizk, Kelsey Bradford, Asim Munawar, Sadhana Kumaravel, Saurabh Goyal, Xin Wang, Luis A. Lastras, Pavan Kapanipathi

Abstract: The resurgence of autonomous agents built using large language models (LLMs) to solve complex real-world tasks has brought increased focus on LLMs' fundamental ability of tool or function calling. At the core of these agents, an LLM must plan, execute, and respond using external tools, APIs, and custom functions. Research on tool calling has gathered momentum, but evaluation benchmarks and dataset… ▽ More The resurgence of autonomous agents built using large language models (LLMs) to solve complex real-world tasks has brought increased focus on LLMs' fundamental ability of tool or function calling. At the core of these agents, an LLM must plan, execute, and respond using external tools, APIs, and custom functions. Research on tool calling has gathered momentum, but evaluation benchmarks and datasets representing the complexity of the tasks have lagged behind. In this work, we focus on one such complexity, nested sequencing, with the goal of extending existing benchmarks and evaluation. Specifically, we present NESTFUL, a benchmark to evaluate LLMs on nested sequences of API calls, i.e., sequences where the output of one API call is passed as input to a subsequent call. NESTFUL contains 1800+ nested sequences where all the function calls are executable. Experimental results on a variety of models show that the best-performing model (GPT-4o) achieves a full sequence match accuracy of 28% and a win-rate of 60%, necessitating a large scope for improvement in the nested sequencing aspect of function calling. Our analysis of these results provides possible future research directions for the community, in addition to a benchmark to track progress. We have released the NESTFUL dataset under the Apache 2.0 license at https://github.com/IBM/NESTFUL. △ Less

Submitted 21 May, 2025; v1 submitted 4 September, 2024; originally announced September 2024.

API-BLEND: A Comprehensive Corpora for Training and Benchmarking API LLMs

Authors: Kinjal Basu, Ibrahim Abdelaziz, Subhajit Chaudhury, Soham Dan, Maxwell Crouse, Asim Munawar, Sadhana Kumaravel, Vinod Muthusamy, Pavan Kapanipathi, Luis A. Lastras

Abstract: 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 cha… ▽ More 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. △ Less

Submitted 20 May, 2024; v1 submitted 23 February, 2024; originally announced February 2024.

Comments: Accepted at ACL'24-main conference

Pointwise Mutual Information Based Metric and Decoding Strategy for Faithful Generation in Document Grounded Dialogs

Authors: Yatin Nandwani, Vineet Kumar, Dinesh Raghu, Sachindra Joshi, Luis A. Lastras

Abstract: A major concern in using deep learning based generative models for document-grounded dialogs is the potential generation of responses that are not \textit{faithful} to the underlying document. Existing automated metrics used for evaluating the faithfulness of response with respect to the grounding document measure the degree of similarity between the generated response and the document's content.… ▽ More A major concern in using deep learning based generative models for document-grounded dialogs is the potential generation of responses that are not \textit{faithful} to the underlying document. Existing automated metrics used for evaluating the faithfulness of response with respect to the grounding document measure the degree of similarity between the generated response and the document's content. However, these automated metrics are far from being well aligned with human judgments. Therefore, to improve the measurement of faithfulness, we propose a new metric that utilizes (Conditional) Point-wise Mutual Information (PMI) between the generated response and the source document, conditioned on the dialogue. PMI quantifies the extent to which the document influences the generated response -- with a higher PMI indicating a more faithful response. We build upon this idea to create a new decoding technique that incorporates PMI into the response generation process to predict more faithful responses. Our experiments on the BEGIN benchmark demonstrate an improved correlation of our metric with human evaluation. We also show that our decoding technique is effective in generating more faithful responses when compared to standard decoding techniques on a set of publicly available document-grounded dialog datasets. △ Less

Submitted 1 December, 2023; v1 submitted 20 May, 2023; originally announced May 2023.

Comments: EMNLP 2023

Towards a Unification of Logic and Information Theory

Authors: Luis A. Lastras, Barry Trager, Jonathan Lenchner, Wojtek Szpankowski, Chai Wah Wu, Mark Squillante, Ron Fagin, Alex Gray

Abstract: Today, the vast majority of the world's digital information is represented using the fundamental assumption, introduced by Claude Shannon in 1948, that ``...the semantic aspects of communication are irrelevant to the engineering problem (of the design of communication systems)...''. Consider, nonetheless, the observation that we often combine a message with other information in order to deduce new… ▽ More Today, the vast majority of the world's digital information is represented using the fundamental assumption, introduced by Claude Shannon in 1948, that ``...the semantic aspects of communication are irrelevant to the engineering problem (of the design of communication systems)...''. Consider, nonetheless, the observation that we often combine a message with other information in order to deduce new facts, thereby expanding the value of such a message. It is noteworthy that to-date, no rigorous theory of communication has been put forth which postulates the existence of deductive capabilities on the receiver's side. The purpose of this paper is to present such a theory. We formally model such deductive capabilities using logic reasoning, and present a rigorous theory which covers the following generic scenario: Alice and Bob each have knowledge of some logic sentence, and they wish to communicate as efficiently as possible with the shared goal that, following their communication, Bob should be able to deduce a particular logic sentence that Alice knows to be true, but that Bob currently cannot prove. Many variants of this general setup are considered in this article; in all cases we are able to provide sharp upper and lower bounds. Our contribution includes the identification of the most fundamental requirements that we place on a logic and associated logical language for all of our results to apply. Practical algorithms that are in some cases asymptotically optimal are provided, and we illustrate the potential practical value of the design of communication systems that incorporate the assumption of deductive capabilities at the receiver using experimental results that suggest significant possible gains compared to classical systems. △ Less

Submitted 12 September, 2025; v1 submitted 25 January, 2023; originally announced January 2023.

DG2: Data Augmentation Through Document Grounded Dialogue Generation

Authors: Qingyang Wu, Song Feng, Derek Chen, Sachindra Joshi, Luis A. Lastras, Zhou Yu

Abstract: Collecting data for training dialog systems can be extremely expensive due to the involvement of human participants and need for extensive annotation. Especially in document-grounded dialog systems, human experts need to carefully read the unstructured documents to answer the users' questions. As a result, existing document-grounded dialog datasets are relatively small-scale and obstruct the effec… ▽ More Collecting data for training dialog systems can be extremely expensive due to the involvement of human participants and need for extensive annotation. Especially in document-grounded dialog systems, human experts need to carefully read the unstructured documents to answer the users' questions. As a result, existing document-grounded dialog datasets are relatively small-scale and obstruct the effective training of dialogue systems. In this paper, we propose an automatic data augmentation technique grounded on documents through a generative dialogue model. The dialogue model consists of a user bot and agent bot that can synthesize diverse dialogues given an input document, which are then used to train a downstream model. When supplementing the original dataset, our method achieves significant improvement over traditional data augmentation methods. We also achieve great performance in the low-resource setting. △ Less

Submitted 15 December, 2021; originally announced December 2021.

doc2dial: A Goal-Oriented Document-Grounded Dialogue Dataset

Authors: Song Feng, Hui Wan, Chulaka Gunasekara, Siva Sankalp Patel, Sachindra Joshi, Luis A. Lastras

Abstract: We introduce doc2dial, a new dataset of goal-oriented dialogues that are grounded in the associated documents. Inspired by how the authors compose documents for guiding end users, we first construct dialogue flows based on the content elements that corresponds to higher-level relations across text sections as well as lower-level relations between discourse units within a section. Then we present t… ▽ More We introduce doc2dial, a new dataset of goal-oriented dialogues that are grounded in the associated documents. Inspired by how the authors compose documents for guiding end users, we first construct dialogue flows based on the content elements that corresponds to higher-level relations across text sections as well as lower-level relations between discourse units within a section. Then we present these dialogue flows to crowd contributors to create conversational utterances. The dataset includes about 4800 annotated conversations with an average of 14 turns that are grounded in over 480 documents from four domains. Compared to the prior document-grounded dialogue datasets, this dataset covers a variety of dialogue scenes in information-seeking conversations. For evaluating the versatility of the dataset, we introduce multiple dialogue modeling tasks and present baseline approaches. △ Less

Submitted 18 November, 2020; v1 submitted 12 November, 2020; originally announced November 2020.

Comments: EMNLP 2020

Lattice Representation Learning

Authors: Luis A. Lastras

Abstract: In this article we introduce theory and algorithms for learning discrete representations that take on a lattice that is embedded in an Euclidean space. Lattice representations possess an interesting combination of properties: a) they can be computed explicitly using lattice quantization, yet they can be learned efficiently using the ideas we introduce in this paper, b) they are highly related to G… ▽ More In this article we introduce theory and algorithms for learning discrete representations that take on a lattice that is embedded in an Euclidean space. Lattice representations possess an interesting combination of properties: a) they can be computed explicitly using lattice quantization, yet they can be learned efficiently using the ideas we introduce in this paper, b) they are highly related to Gaussian Variational Autoencoders, allowing designers familiar with the latter to easily produce discrete representations from their models and c) since lattices satisfy the axioms of a group, their adoption can lead into a way of learning simple algebras for modeling binary operations between objects through symbolic formalisms, yet learn these structures also formally using differentiation techniques. This article will focus on laying the groundwork for exploring and exploiting the first two properties, including a new mathematical result linking expressions used during training and inference time and experimental validation on two popular datasets. △ Less

Submitted 24 June, 2020; originally announced June 2020.

Information Theoretic Lower Bounds on Negative Log Likelihood

Authors: Luis A. Lastras

Abstract: In this article we use rate-distortion theory, a branch of information theory devoted to the problem of lossy compression, to shed light on an important problem in latent variable modeling of data: is there room to improve the model? One way to address this question is to find an upper bound on the probability (equivalently a lower bound on the negative log likelihood) that the model can assign to… ▽ More In this article we use rate-distortion theory, a branch of information theory devoted to the problem of lossy compression, to shed light on an important problem in latent variable modeling of data: is there room to improve the model? One way to address this question is to find an upper bound on the probability (equivalently a lower bound on the negative log likelihood) that the model can assign to some data as one varies the prior and/or the likelihood function in a latent variable model. The core of our contribution is to formally show that the problem of optimizing priors in latent variable models is exactly an instance of the variational optimization problem that information theorists solve when computing rate-distortion functions, and then to use this to derive a lower bound on negative log likelihood. Moreover, we will show that if changing the prior can improve the log likelihood, then there is a way to change the likelihood function instead and attain the same log likelihood, and thus rate-distortion theory is of relevance to both optimizing priors as well as optimizing likelihood functions. We will experimentally argue for the usefulness of quantities derived from rate-distortion theory in latent variable modeling by applying them to a problem in image modeling. △ Less

Submitted 12 April, 2019; originally announced April 2019.

A Proximity Measure using Blink Model

Authors: Haifeng Qian, Hui Wan, Mark N. Wegman, Luis A. Lastras, Ruchir Puri

Abstract: This paper proposes a new graph proximity measure. This measure is a derivative of network reliability. By analyzing its properties and comparing it against other proximity measures through graph examples, we demonstrate that it is more consistent with human intuition than competitors. A new deterministic algorithm is developed to approximate this measure with practical complexity. Empirical evalu… ▽ More This paper proposes a new graph proximity measure. This measure is a derivative of network reliability. By analyzing its properties and comparing it against other proximity measures through graph examples, we demonstrate that it is more consistent with human intuition than competitors. A new deterministic algorithm is developed to approximate this measure with practical complexity. Empirical evaluation by two link prediction benchmarks, one in coauthorship networks and one in Wikipedia, shows promising results. For example, a single parameterization of this measure achieves accuracies that are 14-35% above the best accuracy for each graph of all predictors reported in the 2007 Liben-Nowell and Kleinberg survey. △ Less

Submitted 21 December, 2016; originally announced December 2016.