Harsh Jhamtani

Kumar Shridhar, Harsh Jhamtani, Hao Fang et al. · microsoft-research

Large language models (LLMs) can improve their accuracy on various tasks through iteratively refining and revising their output based on feedback. We observe that these revisions can introduce errors, in which case it is better to roll back to a previous result. Further, revisions are typically homogeneous: they use the same reasoning method that produced the initial answer, which may not correct errors. To enable exploration in this space, we present SCREWS, a modular framework for reasoning with revisions. It is comprised of three main modules: Sampling, Conditional Resampling, and Selection, each consisting of sub-modules that can be hand-selected per task. We show that SCREWS not only unifies several previous approaches under a common framework, but also reveals several novel strategies for identifying improved reasoning chains. We evaluate our framework with state-of-the-art LLMs (ChatGPT and GPT-4) on a diverse set of reasoning tasks and uncover useful new reasoning strategies for each: arithmetic word problems, multi-hop question answering, and code debugging. Heterogeneous revision strategies prove to be important, as does selection between original and revised candidates.

Hao Fang, Anusha Balakrishnan, Harsh Jhamtani et al. · microsoft-research, mit

In a real-world dialogue system, generated text must be truthful and informative while remaining fluent and adhering to a prescribed style. Satisfying these constraints simultaneously is difficult for the two predominant paradigms in language generation: neural language modeling and rule-based generation. We describe a hybrid architecture for dialogue response generation that combines the strengths of both paradigms. The first component of this architecture is a rule-based content selection model defined using a new formal framework called dataflow transduction, which uses declarative rules to transduce a dialogue agent's actions and their results (represented as dataflow graphs) into context-free grammars representing the space of contextually acceptable responses. The second component is a constrained decoding procedure that uses these grammars to constrain the output of a neural language model, which selects fluent utterances. Our experiments show that this system outperforms both rule-based and learned approaches in human evaluations of fluency, relevance, and truthfulness.

Prakhar Gupta, Harsh Jhamtani, Jeffrey P. Bigham · cmu, microsoft-research

Target-guided response generation enables dialogue systems to smoothly transition a conversation from a dialogue context toward a target sentence. Such control is useful for designing dialogue systems that direct a conversation toward specific goals, such as creating non-obtrusive recommendations or introducing new topics in the conversation. In this paper, we introduce a new technique for target-guided response generation, which first finds a bridging path of commonsense knowledge concepts between the source and the target, and then uses the identified bridging path to generate transition responses. Additionally, we propose techniques to re-purpose existing dialogue datasets for target-guided generation. Experiments reveal that the proposed techniques outperform various baselines on this task. Finally, we observe that the existing automated metrics for this task correlate poorly with human judgement ratings. We propose a novel evaluation metric that we demonstrate is more reliable for target-guided response evaluation. Our work generally enables dialogue system designers to exercise more control over the conversations that their systems produce.

Bodhisattwa Prasad Majumder, Harsh Jhamtani, Taylor Berg-Kirkpatrick et al. · microsoft-research

A limitation of current neural dialog models is that they tend to suffer from a lack of specificity and informativeness in generated responses, primarily due to dependence on training data that covers a limited variety of scenarios and conveys limited knowledge. One way to alleviate this issue is to extract relevant knowledge from external sources at decoding time and incorporate it into the dialog response. In this paper, we propose a post-hoc knowledge-injection technique where we first retrieve a diverse set of relevant knowledge snippets conditioned on both the dialog history and an initial response from an existing dialog model. We construct multiple candidate responses, individually injecting each retrieved snippet into the initial response using a gradient-based decoding method, and then select the final response with an unsupervised ranking step. Our experiments in goal-oriented and knowledge-grounded dialog settings demonstrate that human annotators judge the outputs from the proposed method to be more engaging and informative compared to responses from prior dialog systems. We further show that knowledge-augmentation promotes success in achieving conversational goals in both experimental settings.

Nathaniel Weir, Ryan Thomas, Randolph D'Amore et al. · microsoft-research

We introduce a language generation task grounded in a popular video game environment. KNUDGE (KNowledge Constrained User-NPC Dialogue GEneration) requires models to produce trees of dialogue between video game characters that accurately reflect quest and entity specifications stated in natural language. KNUDGE is constructed from side quest dialogues drawn directly from game data of Obsidian Entertainment's The Outer Worlds, leading to real-world complexities in generation: (1) dialogues are branching trees as opposed to linear chains of utterances; (2) utterances must remain faithful to the game lore -- character personas, backstories, and entity relationships; and (3) a dialogue must accurately reveal new quest details to the human player. We report results for a set of neural generation models using supervised and in-context learning techniques; we find competent performance but room for future work addressing the challenges of creating realistic, game-quality dialogues.

Sedrick Scott Keh, Kevin Lu, Varun Gangal et al. · amazon-science, cmu

A personification is a figure of speech that endows inanimate entities with properties and actions typically seen as requiring animacy. In this paper, we explore the task of personification generation. To this end, we propose PINEAPPLE: Personifying INanimate Entities by Acquiring Parallel Personification data for Learning Enhanced generation. We curate a corpus of personifications called PersonifCorp, together with automatically generated de-personified literalizations of these personifications. We demonstrate the usefulness of this parallel corpus by training a seq2seq model to personify a given literal input. Both automatic and human evaluations show that fine-tuning with PersonifCorp leads to significant gains in personification-related qualities such as animacy and interestingness. A detailed qualitative analysis also highlights key strengths and imperfections of PINEAPPLE over baselines, demonstrating a strong ability to generate diverse and creative personifications that enhance the overall appeal of a sentence.

Nikita Moghe, Patrick Xia, Jacob Andreas et al. · microsoft-research

Users of natural language interfaces, generally powered by Large Language Models (LLMs),often must repeat their preferences each time they make a similar request. We describe an approach to LLM-based dialogue modeling in which persistent user constraints and preferences -- collectively termed standing instructions -- as additional context for such interfaces. For example, when a user states "I'm hungry", a previously expressed preference for Persian food can be automatically added to the LLM prompt, influencing the search for relevant restaurants. We develop NLSI, a language-to-program dataset consisting of over 2.4K dialogues spanning 17 domains, where each dialogue is paired with a user profile (a set of users specific standing instructions) and corresponding structured representations (API calls). A key challenge in NLSI is to identify which subset of the standing instructions is applicable to a given dialogue. NLSI contains diverse phenomena, from simple preferences to interdependent instructions such as triggering a hotel search whenever the user is booking tickets to an event. We conduct experiments on NLSI using prompting with large language models and various retrieval approaches, achieving a maximum of 44.7% exact match on API prediction. Our results demonstrate the challenges in identifying the relevant standing instructions and their interpretation into API calls.

Abhinav Joshi, Shaswati Saha, Divyaksh Shukla et al. · microsoft-research

Large Language Models (LLMs) have shown to be a great success in a wide range of applications ranging from regular NLP-based use cases to AI agents. LLMs have been trained on a vast corpus of texts from various sources; despite the best efforts during the data pre-processing stage while training the LLMs, they may pick some undesirable information such as personally identifiable information (PII). Consequently, in recent times research in the area of Machine Unlearning (MUL) has become active, the main idea is to force LLMs to forget (unlearn) certain information (e.g., PII) without suffering from performance loss on regular tasks. In this work, we examine the robustness of the existing MUL techniques for their ability to enable leakage-proof forgetting in LLMs. In particular, we examine the effect of data transformation on forgetting, i.e., is an unlearned LLM able to recall forgotten information if there is a change in the format of the input? Our findings on the TOFU dataset highlight the necessity of using diverse data formats to quantify unlearning in LLMs more reliably.

Harsh Jhamtani, Jacob Andreas, Benjamin Van Durme · microsoft-research

This paper introduces PeopleJoin, a benchmark for evaluating LM-mediated collaborative problem solving. Given a user request, PeopleJoin agents must identify teammates who might be able to assist, converse with these teammates to gather information, and finally compile a useful answer or summary for the original user. PeopleJoin comprises two evaluation domains: PeopleJoin-QA, focused on questions about tabular data, and PeopleJoin-DocCreation, focused on document creation tasks. The two domains are adapted from existing NLP benchmarks for database question answering and multi-document summarization; here, however, the information needed to complete these tasks is distributed across synthetic ``organizations'' of 2--20 users, simulating natural multi-user collaboration scenarios. We implemented several popular LM agent architectures, evaluating their accuracy and efficiency at completing tasks, and highlight new research questions that can be studied using PeopleJoin.

Michael Y. Hu, Benjamin Van Durme, Jacob Andreas et al. · microsoft-research

Language model (LM) agents deployed in novel environments often exhibit poor sample efficiency when learning from sequential interactions. This significantly hinders the usefulness of such agents in environments where interaction is costly (for example, when they interact with humans or reset physical systems). While a number of existing LM agent architectures incorporate various mechanisms for experience storage and reflection, they make limited use of LMs' abilities to directly generate or reason about full counterfactual trajectories. We introduce ECHO (Experience Consolidation via Hindsight Optimization), a prompting framework that adapts hindsight experience replay from reinforcement learning for language model agents. ECHO generates optimized trajectories for alternative goals that could have been achieved during failed attempts, effectively creating synthetic positive examples from unsuccessful interactions. Our approach consists of two components: a hindsight rule that uses the language model itself to identify relevant subgoals and generate optimized trajectories, and an update rule that maintains compressed trajectory representations in memory. We evaluate ECHO on stateful versions of XMiniGrid, a text-based navigation and planning benchmark, and PeopleJoinQA, a collaborative information-gathering enterprise simulation. Across both domains, ECHO outperforms vanilla language agent baselines by up to 80%; in XMiniGrid, it also outperforms a number of sophisticated agent architectures including Reflexion and AWM, demonstrating faster adaptation to novel environments through more effective utilization of past experiences.

Yunmo Chen, Tongfei Chen, Harsh Jhamtani et al.

We introduce iterative retrieval, a novel framework that empowers retrievers to make iterative decisions through policy optimization. Finding an optimal portfolio of retrieved items is a combinatorial optimization problem, generally considered NP-hard. This approach provides a learned approximation to such a solution, meeting specific task requirements under a given family of large language models (LLMs). We propose a training procedure based on reinforcement learning, incorporating feedback from LLMs. We instantiate an iterative retriever for composing in-context learning (ICL) exemplars and apply it to various semantic parsing tasks that demand synthesized programs as outputs. By adding only 4M additional parameters for state encoding, we convert an off-the-shelf dense retriever into a stateful iterative retriever, outperforming previous methods in selecting ICL exemplars on semantic parsing datasets such as CalFlow, TreeDST, and MTOP. Additionally, the trained iterative retriever generalizes across different inference LLMs beyond the one used during training.

Harsh Jhamtani, Hao Fang, Patrick Xia et al.

Designing natural language interfaces has historically required collecting supervised data to translate user requests into carefully designed intent representations. This requires enumerating and labeling a long tail of user requests, which is challenging. At the same time, large language models (LLMs) encode knowledge about goals and plans that can help conversational assistants interpret user requests requiring numerous steps to complete. We introduce an approach to handle complex-intent-bearing utterances from a user via a process of hierarchical natural language decomposition and interpretation. Our approach uses a pre-trained language model to decompose a complex utterance into a sequence of simpler natural language steps and interprets each step using the language-to-program model designed for the interface. To test our approach, we collect and release DeCU -- a new NL-to-program benchmark to evaluate Decomposition of Complex Utterances. Experiments show that the proposed approach enables the interpretation of complex utterances with almost no complex training data, while outperforming standard few-shot prompting approaches.

Harsh Jhamtani, Taylor Berg-Kirkpatrick

In this paper, we explore the task of automatically generating natural language descriptions of salient patterns in a time series, such as stock prices of a company over a week. A model for this task should be able to extract high-level patterns such as presence of a peak or a dip. While typical contemporary neural models with attention mechanisms can generate fluent output descriptions for this task, they often generate factually incorrect descriptions. We propose a computational model with a truth-conditional architecture which first runs small learned programs on the input time series, then identifies the programs/patterns which hold true for the given input, and finally conditions on only the chosen valid program (rather than the input time series) to generate the output text description. A program in our model is constructed from modules, which are small neural networks that are designed to capture numerical patterns and temporal information. The modules are shared across multiple programs, enabling compositionality as well as efficient learning of module parameters. The modules, as well as the composition of the modules, are unobserved in data, and we learn them in an end-to-end fashion with the only training signal coming from the accompanying natural language text descriptions. We find that the proposed model is able to generate high-precision captions even though we consider a small and simple space of module types.

Harsh Jhamtani, Varun Gangal, Eduard Hovy et al.

Humans often employ figurative language use in communication, including during interactions with dialog systems. Thus, it is important for real-world dialog systems to be able to handle popular figurative language constructs like metaphor and simile. In this work, we analyze the performance of existing dialog models in situations where the input dialog context exhibits use of figurative language. We observe large gaps in handling of figurative language when evaluating the models on two open domain dialog datasets. When faced with dialog contexts consisting of figurative language, some models show very large drops in performance compared to contexts without figurative language. We encourage future research in dialog modeling to separately analyze and report results on figurative language in order to better test model capabilities relevant to real-world use. Finally, we propose lightweight solutions to help existing models become more robust to figurative language by simply using an external resource to translate figurative language to literal (non-figurative) forms while preserving the meaning to the best extent possible.

Bodhisattwa Prasad Majumder, Taylor Berg-Kirkpatrick, Julian McAuley et al.

Humans often refer to personal narratives, life experiences, and events to make a conversation more engaging and rich. While persona-grounded dialog models are able to generate responses that follow a given persona, they often miss out on stating detailed experiences or events related to a persona, often leaving conversations shallow and dull. In this work, we equip dialog models with 'background stories' related to a persona by leveraging fictional narratives from existing story datasets (e.g. ROCStories). Since current dialog datasets do not contain such narratives as responses, we perform an unsupervised adaptation of a retrieved story for generating a dialog response using a gradient-based rewriting technique. Our proposed method encourages the generated response to be fluent (i.e., highly likely) with the dialog history, minimally different from the retrieved story to preserve event ordering and consistent with the original persona. We demonstrate that our method can generate responses that are more diverse, and are rated more engaging and human-like by human evaluators, compared to outputs from existing dialog models.

Varun Gangal, Harsh Jhamtani, Eduard Hovy et al.

Multiple different responses are often plausible for a given open domain dialog context. Prior work has shown the importance of having multiple valid reference responses for meaningful and robust automated evaluations. In such cases, common practice has been to collect more human written references. However, such collection can be expensive, time consuming, and not easily scalable. Instead, we propose a novel technique for automatically expanding a human generated reference to a set of candidate references. We fetch plausible references from knowledge sources, and adapt them so that they are more fluent in context of the dialog instance in question. More specifically, we use (1) a commonsense knowledge base to elicit a large number of plausible reactions given the dialog history (2) relevant instances retrieved from dialog corpus, using similar past as well as future contexts. We demonstrate that our automatically expanded reference sets lead to large improvements in correlations of automated metrics with human ratings of system outputs for DailyDialog dataset.

Sebastian Gehrmann, Tosin Adewumi, Karmanya Aggarwal et al.

We introduce GEM, a living benchmark for natural language Generation (NLG), its Evaluation, and Metrics. Measuring progress in NLG relies on a constantly evolving ecosystem of automated metrics, datasets, and human evaluation standards. Due to this moving target, new models often still evaluate on divergent anglo-centric corpora with well-established, but flawed, metrics. This disconnect makes it challenging to identify the limitations of current models and opportunities for progress. Addressing this limitation, GEM provides an environment in which models can easily be applied to a wide set of tasks and in which evaluation strategies can be tested. Regular updates to the benchmark will help NLG research become more multilingual and evolve the challenge alongside models. This paper serves as the description of the data for which we are organizing a shared task at our ACL 2021 Workshop and to which we invite the entire NLG community to participate.

Harsh Jhamtani, Peter Clark

Despite the rapid progress in multihop question-answering (QA), models still have trouble explaining why an answer is correct, with limited explanation training data available to learn from. To address this, we introduce three explanation datasets in which explanations formed from corpus facts are annotated. Our first dataset, eQASC, contains over 98K explanation annotations for the multihop question answering dataset QASC, and is the first that annotates multiple candidate explanations for each answer. The second dataset eQASC-perturbed is constructed by crowd-sourcing perturbations (while preserving their validity) of a subset of explanations in QASC, to test consistency and generalization of explanation prediction models. The third dataset eOBQA is constructed by adding explanation annotations to the OBQA dataset to test generalization of models trained on eQASC. We show that this data can be used to significantly improve explanation quality (+14% absolute F1 over a strong retrieval baseline) using a BERT-based classifier, but still behind the upper bound, offering a new challenge for future research. We also explore a delexicalized chain representation in which repeated noun phrases are replaced by variables, thus turning them into generalized reasoning chains (for example: "X is a Y" AND "Y has Z" IMPLIES "X has Z"). We find that generalized chains maintain performance while also being more robust to certain perturbations.

Harsh Jhamtani, Taylor Berg-Kirkpatrick

Past work on story generation has demonstrated the usefulness of conditioning on a generation plan to generate coherent stories. However, these approaches have used heuristics or off-the-shelf models to first tag training stories with the desired type of plan, and then train generation models in a supervised fashion. In this paper, we propose a deep latent variable model that first samples a sequence of anchor words, one per sentence in the story, as part of its generative process. During training, our model treats the sequence of anchor words as a latent variable and attempts to induce anchoring sequences that help guide generation in an unsupervised fashion. We conduct experiments with several types of sentence decoder distributions: left-to-right and non-monotonic, with different degrees of restriction. Further, since we use amortized variational inference to train our model, we introduce two corresponding types of inference network for predicting the posterior on anchor words. We conduct human evaluations which demonstrate that the stories produced by our model are rated better in comparison with baselines which do not consider story plans, and are similar or better in quality relative to baselines which use external supervision for plans. Additionally, the proposed model gets favorable scores when evaluated on perplexity, diversity, and control of story via discrete plan.

Bodhisattwa Prasad Majumder, Harsh Jhamtani, Taylor Berg-Kirkpatrick et al.

Existing persona-grounded dialog models often fail to capture simple implications of given persona descriptions, something which humans are able to do seamlessly. For example, state-of-the-art models cannot infer that interest in hiking might imply love for nature or longing for a break. In this paper, we propose to expand available persona sentences using existing commonsense knowledge bases and paraphrasing resources to imbue dialog models with access to an expanded and richer set of persona descriptions. Additionally, we introduce fine-grained grounding on personas by encouraging the model to make a discrete choice among persona sentences while synthesizing a dialog response. Since such a choice is not observed in the data, we model it using a discrete latent random variable and use variational learning to sample from hundreds of persona expansions. Our model outperforms competitive baselines on the PersonaChat dataset in terms of dialog quality and diversity while achieving persona-consistent and controllable dialog generation.

Harsh Jhamtani, Sanket Vaibhav Mehta, Jaime Carbonell et al.

Existing recurrent neural language models often fail to capture higher-level structure present in text: for example, rhyming patterns present in poetry. Much prior work on poetry generation uses manually defined constraints which are satisfied during decoding using either specialized decoding procedures or rejection sampling. The rhyming constraints themselves are typically not learned by the generator. We propose an alternate approach that uses a structured discriminator to learn a poetry generator that directly captures rhyming constraints in a generative adversarial setup. By causing the discriminator to compare poems based only on a learned similarity matrix of pairs of line ending words, the proposed approach is able to successfully learn rhyming patterns in two different English poetry datasets (Sonnet and Limerick) without explicitly being provided with any phonetic information.

Harsh Jhamtani, Taylor Berg-Kirkpatrick

In this paper, we introduce the task of automatically generating text to describe the differences between two similar images. We collect a new dataset by crowd-sourcing difference descriptions for pairs of image frames extracted from video-surveillance footage. Annotators were asked to succinctly describe all the differences in a short paragraph. As a result, our novel dataset provides an opportunity to explore models that align language and vision, and capture visual salience. The dataset may also be a useful benchmark for coherent multi-sentence generation. We perform a firstpass visual analysis that exposes clusters of differing pixels as a proxy for object-level differences. We propose a model that captures visual salience by using a latent variable to align clusters of differing pixels with output sentences. We find that, for both single-sentence generation and as well as multi-sentence generation, the proposed model outperforms the models that use attention alone.

Anant Subramanian, Danish Pruthi, Harsh Jhamtani et al.

Prediction without justification has limited utility. Much of the success of neural models can be attributed to their ability to learn rich, dense and expressive representations. While these representations capture the underlying complexity and latent trends in the data, they are far from being interpretable. We propose a novel variant of denoising k-sparse autoencoders that generates highly efficient and interpretable distributed word representations (word embeddings), beginning with existing word representations from state-of-the-art methods like GloVe and word2vec. Through large scale human evaluation, we report that our resulting word embedddings are much more interpretable than the original GloVe and word2vec embeddings. Moreover, our embeddings outperform existing popular word embeddings on a diverse suite of benchmark downstream tasks.

Varun Gangal, Harsh Jhamtani, Graham Neubig et al.

Portmanteaus are a word formation phenomenon where two words are combined to form a new word. We propose character-level neural sequence-to-sequence (S2S) methods for the task of portmanteau generation that are end-to-end-trainable, language independent, and do not explicitly use additional phonetic information. We propose a noisy-channel-style model, which allows for the incorporation of unsupervised word lists, improving performance over a standard source-to-target model. This model is made possible by an exhaustive candidate generation strategy specifically enabled by the features of the portmanteau task. Experiments find our approach superior to a state-of-the-art FST-based baseline with respect to ground truth accuracy and human evaluation.

Harsh Jhamtani, Varun Gangal, Eduard Hovy et al.

Variations in writing styles are commonly used to adapt the content to a specific context, audience, or purpose. However, applying stylistic variations is still by and large a manual process, and there have been little efforts towards automating it. In this paper we explore automated methods to transform text from modern English to Shakespearean English using an end to end trainable neural model with pointers to enable copy action. To tackle limited amount of parallel data, we pre-train embeddings of words by leveraging external dictionaries mapping Shakespearean words to modern English words as well as additional text. Our methods are able to get a BLEU score of 31+, an improvement of ~6 points above the strongest baseline. We publicly release our code to foster further research in this area.

Gaurush Hiranandani, Pranav Maneriker, Harsh Jhamtani

Providing appealing brand names to newly launched products, newly formed companies or for renaming existing companies is highly important as it can play a crucial role in deciding its success or failure. In this work, we propose a computational method to generate appealing brand names based on the description of such entities. We use quantitative scores for readability, pronounceability, memorability and uniqueness of the generated names to rank order them. A set of diverse appealing names is recommended to the user for the brand naming task. Experimental results show that the names generated by our approach are more appealing than names which prior approaches and recruited humans could come up.