Razvan Bunescu

Erfan Al-Hossami, Razvan Bunescu, Justin Smith et al.

When employing the Socratic method of teaching, instructors guide students toward solving a problem on their own rather than providing the solution directly. While this strategy can substantially improve learning outcomes, it is usually time-consuming and cognitively demanding. Automated Socratic conversational agents can augment human instruction and provide the necessary scale, however their development is hampered by the lack of suitable data for training and evaluation. In this paper, we introduce a manually created dataset of multi-turn Socratic advice that is aimed at helping a novice programmer fix buggy solutions to simple computational problems. The dataset is then used for benchmarking the Socratic debugging abilities of a number of language models, ranging from fine-tuning the instruction-based text-to-text transformer Flan-T5 to zero-shot and chain of thought prompting of the much larger GPT-4. The code and datasets are made freely available for research at the link below. https://github.com/taisazero/socratic-debugging-benchmark

Tonmoy Hasan, Razvan Bunescu

A recommender system that optimizes its recommendations solely to fit a user's history of ratings for consumed items can create a filter bubble, wherein the user does not get to experience items from novel, unseen categories. One approach to mitigate this undesired behavior is to recommend items with high potential for serendipity, namely surprising items that are likely to be highly rated. In this paper, we propose a content-based formulation of serendipity that is rooted in Bayesian surprise and use it to measure the serendipity of items after they are consumed and rated by the user. When coupled with a collaborative-filtering component that identifies similar users, this enables recommending items with high potential for serendipity. To facilitate the evaluation of topic-level models for surprise and serendipity, we introduce a dataset of book reading histories extracted from Goodreads, containing over 26 thousand users and close to 1.3 million books, where we manually annotate 449 books read by 4 users in terms of their time-dependent, topic-level surprise. Experimental evaluations show that models that use Bayesian surprise correlate much better with the manual annotations of topic-level surprise than distance-based heuristics, and also obtain better serendipitous item recommendation performance.

Quintin Fettes, Avinash Karanth, Razvan Bunescu et al.

Many cloud applications are migrated from the monolithic model to a microservices framework in which hundreds of loosely-coupled microservices run concurrently, with significant benefits in terms of scalability, rapid development, modularity, and isolation. However, dependencies among microservices with uneven execution time may result in longer queues, idle resources, or Quality-of-Service (QoS) violations. In this paper we introduce Reclaimer, a deep reinforcement learning model that adapts to runtime changes in the number and behavior of microservices in order to minimize CPU core allocation while meeting QoS requirements. When evaluated with two benchmark microservice-based applications, Reclaimer reduces the mean CPU core allocation by 38.4% to 74.4% relative to the industry-standard scaling solution, and by 27.5% to 58.1% relative to a current state-of-the art method.

Erfan Al-Hossami, Razvan Bunescu

In Socratic debugging, instructors guide students towards identifying and fixing a bug on their own, instead of providing the bug fix directly. Most novice programmer bugs are caused by programming misconceptions, namely false beliefs about a programming concept. In this context, Socratic debugging can be formulated as a guided Reasoning Trajectory (RT) leading to a statement about the program behavior that contradicts the bug-causing misconception. Upon reaching this statement, the ensuing cognitive dissonance leads the student to first identify and then update their false belief. In this paper, we introduce the task of reasoning trajectory generation, together with a dataset of debugging problems manually annotated with RTs. We then describe LLM-based solutions for generating RTs and Socratic conversations that are anchored on them. A large-scale LLM-as-judge evaluation shows that frontier models can generate up to 91% correct reasoning trajectories and 98.7% valid conversation turns.

Smita Nannaware, Erfan Al-Hossami, Razvan Bunescu

A restaurant dinner may become a memorable experience due to an unexpected aspect enjoyed by the customer, such as an origami-making station in the waiting area. If aspects that are atypical for a restaurant experience were known in advance, they could be leveraged to make recommendations that have the potential to engender serendipitous experiences, further increasing user satisfaction. Although relatively rare, whenever encountered, atypical aspects often end up being mentioned in reviews due to their memorable quality. Correspondingly, in this paper we introduce the task of detecting atypical aspects in customer reviews. To facilitate the development of extraction models, we manually annotate benchmark datasets of reviews in three domains - restaurants, hotels, and hair salons, which we use to evaluate a number of language models, ranging from fine-tuning the instruction-based text-to-text transformer Flan-T5 to zero-shot and few-shot prompting of GPT-3.5.

Ramit Aditya, Razvan Bunescu, Smita Nannaware et al.

A restaurant dinner or a hotel stay may lead to memorable experiences when guests encounter unexpected aspects that also match their interests. For example, an origami-making station in the waiting area of a restaurant may be both surprising and enjoyable for a customer who is passionate about paper crafts. Similarly, an exhibit of 18th century harpsichords would be atypical for a hotel lobby and likely pique the interest of a guest who has a passion for Baroque music. Motivated by this insight, in this paper we introduce the new task of engineering serendipity through recommendations of items with atypical aspects. We describe an LLM-based system pipeline that extracts atypical aspects from item reviews, then estimates and aggregates their user-specific utility in a measure of serendipity potential that is used to rerank a list of items recommended to the user. To facilitate system development and evaluation, we introduce a dataset of Yelp reviews that are manually annotated with atypical aspects and a dataset of artificially generated user profiles, together with crowdsourced annotations of user-aspect utility values. Furthermore, we introduce a custom procedure for dynamic selection of in-context learning examples, which is shown to improve LLM-based judgments of atypicality and utility. Experimental evaluations show that serendipity-based rankings generated by the system are highly correlated with ground truth rankings for which serendipity scores are computed from manual annotations of atypical aspects and their user-dependent utility. Overall, we hope that the new recommendation task and the associated system presented in this paper catalyze further research into recommendation approaches that go beyond accuracy in their pursuit of enhanced user satisfaction. The datasets and the code are made publicly available at https://github.com/ramituncc49er/ATARS .

Erik Saule, Kalpathi Subramanian, Razvan Bunescu

Professional societies often publish curriculum guidelines to help programs align their content to international standards. In Computer Science, the primary standard is published by ACM and IEEE and provide detailed guidelines for what should be and could be included in a Computer Science program. While very helpful, it remains difficult for program administrators to assess how much of the guidelines is being covered by a CS program. This is in particular due to the extensiveness of the guidelines, containing thousands of individual items. As such, it is time consuming and cognitively demanding to audit every course to confidently mark everything that is actually being covered. Our preliminary work indicated that it takes about a day of work per course. In this work, we propose using Natural Language Processing techniques to accelerate the process. We explore two kinds of techniques, the first relying on traditional tools for parsing, tagging, and embeddings, while the second leverages the power of Large Language Models. We evaluate the application of these techniques to classify a corpus of pedagogical materials and show that we can meaningfully classify documents automatically.

Erfan Al-Hossami, Razvan Bunescu

When learning to code, students often develop misconceptions about various programming language concepts. These can not only lead to bugs or inefficient code, but also slow down the learning of related concepts. In this paper, we introduce McMining, the task of mining programming misconceptions from samples of code from a student. To enable the training and evaluation of McMining systems, we develop an extensible benchmark dataset of misconceptions together with a large set of code samples where these misconceptions are manifested. We then introduce two LLM-based McMiner approaches and through extensive evaluations show that models from the Gemini, Claude, and GPT families are effective at discovering misconceptions in student code.

Youssef A. Ait Alama, Sampada Sakpal, Ke Wang et al.

Hardware failures are a growing challenge for machine learning accelerators, many of which are based on systolic arrays. When a permanent hardware failure occurs in a systolic array, existing solutions include localizing and isolating the faulty processing element (PE), using a redundant PE for re-execution, or in some extreme cases decommissioning the entire accelerator for further investigation. In this paper, we propose novel algorithmic approaches that mitigate permanent hardware faults in neural network (NN) accelerators by uniquely integrating the behavior of the faulty component instead of bypassing it. In doing so, we aim for a more sustainable use of the accelerator where faulty hardware is neither bypassed nor discarded, instead being given a second life. We first introduce a CUDA-accelerated systolic array simulator in PyTorch, which enabled us to quantify the impact of permanent faults appearing on links connecting two PEs or in weight registers, where one bit is stuck at 0 or 1 in the float32, float16, or bfloat16 representation. We then propose several algorithmic mitigation techniques for a subset of stuck-at faults, such as Invertible Scaling or Shifting of activations and weights, or fine tuning with the faulty behavior. Notably, the proposed techniques do not require any hardware modification, instead relying on existing components of widely used systolic array based accelerators, such as normalization, activation, and storage units. Extensive experimental evaluations using fully connected and convolutional NNs trained on MNIST, CIFAR-10 and ImageNet show that the proposed fault-tolerant approach matches or gets very close to the original fault-free accuracy.

Jeremy Beauchamp, Razvan Bunescu, Cindy Marling et al.

To avoid serious diabetic complications, people with type 1 diabetes must keep their blood glucose levels (BGLs) as close to normal as possible. Insulin dosages and carbohydrate consumption are important considerations in managing BGLs. Since the 1960s, models have been developed to forecast blood glucose levels based on the history of BGLs, insulin dosages, carbohydrate intake, and other physiological and lifestyle factors. Such predictions can be used to alert people of impending unsafe BGLs or to control insulin flow in an artificial pancreas. In past work, we have introduced an LSTM-based approach to blood glucose level prediction aimed at "what if" scenarios, in which people could enter foods they might eat or insulin amounts they might take and then see the effect on future BGLs. In this work, we invert the "what-if" scenario and introduce a similar architecture based on chaining two LSTMs that can be trained to make either insulin or carbohydrate recommendations aimed at reaching a desired BG level in the future. Leveraging a recent state-of-the-art model for time series forecasting, we then derive a novel architecture for the same recommendation task, in which the two LSTM chain is used as a repeating block inside a deep residual architecture. Experimental evaluations using real patient data from the OhioT1DM dataset show that the new integrated architecture compares favorably with the previous LSTM-based approach, substantially outperforming the baselines. The promising results suggest that this novel approach could potentially be of practical use to people with type 1 diabetes for self-management of BGLs.

Mike Chen, Razvan Bunescu

We introduce the task of changing the narrative point of view, where characters are assigned a narrative perspective that is different from the one originally used by the writer. The resulting shift in the narrative point of view alters the reading experience and can be used as a tool in fiction writing or to generate types of text ranging from educational to self-help and self-diagnosis. We introduce a benchmark dataset containing a wide range of types of narratives annotated with changes in point of view from deictic (first or second person) to anaphoric (third person) and describe a pipeline for processing raw text that relies on a neural architecture for mention selection. Evaluations on the new benchmark dataset show that the proposed architecture substantially outperforms the baselines by generating mentions that are less ambiguous and more natural.

Kaiyu Shen, Razvan Bunescu, Sarah E. Wyatt

The study of biological processes can greatly benefit from tools that automatically predict gene functions or directly cluster genes based on shared functionality. Existing data mining methods predict protein functionality by exploiting data obtained from high-throughput experiments or meta-scale information from public databases. Most existing prediction tools are targeted at predicting protein functions that are described in the gene ontology (GO). However, in many cases biologists wish to discover functionally related genes for which GO terms are inadequate. In this paper, we introduce a rich set of features and use them in conjunction with semisupervised learning approaches in order to expand an initial set of seed genes to a larger cluster of functionally related genes. Among all the semi-supervised methods that were evaluated, the framework of learning with positive and unlabeled examples (LPU) is shown to be especially appropriate for mining functionally related genes. When evaluated on experimentally validated benchmark data, the LPU approaches1 significantly outperform a standard supervised learning algorithm as well as an established state-of-the-art method. Given an initial set of seed genes, our best performing approach could be used to mine functionally related genes in a wide range of organisms.

Patrick Gray, Razvan Bunescu

Music is often experienced as a progression of concurrent streams of notes, or voices. The degree to which this happens depends on the position along a voice-leading continuum, ranging from monophonic, to homophonic, to polyphonic, which complicates the design of automatic voice separation models. We address this continuum by defining voice separation as the task of decomposing music into streams that exhibit both a high degree of external perceptual separation from the other streams and a high degree of internal perceptual consistency. The proposed voice separation task allows for a voice to diverge to multiple voices and also for multiple voices to converge to the same voice. Equipped with this flexible task definition, we manually annotated a corpus of popular music and used it to train neural networks that assign notes to voices either separately for each note in a chord (note-level), or jointly to all notes in a chord (chord-level). The trained neural models greedily assign notes to voices in a left to right traversal of the input chord sequence, using a diverse set of perceptually informed input features. When evaluated on the extraction of consecutive within voice note pairs, both models surpass a strong baseline based on an iterative application of an envelope extraction function, with the chord-level model consistently edging out the note-level model. The two models are also shown to outperform previous approaches on separating the voices in Bach music.

Xianlong Zeng, Soheil Moosavinasab, En-Ju D Lin et al.

Efficient representation of patients is very important in the healthcare domain and can help with many tasks such as medical risk prediction. Many existing methods, such as diagnostic Cost Groups (DCG), rely on expert knowledge to build patient representation from medical data, which is resource consuming and non-scalable. Unsupervised machine learning algorithms are a good choice for automating the representation learning process. However, there is very little research focusing on onpatient-level representation learning directly from medical claims. In this paper, weproposed a novel patient vector learning architecture that learns high quality,fixed-length patient representation from claims data. We conducted several experiments to test the quality of our learned representation, and the empirical results show that our learned patient vectors are superior to vectors learned through other methods including a popular commercial model. Lastly, we provide potential clinical interpretation for using our representation on predictive tasks, as interpretability is vital in the healthcare domain

Charles Chen, Ruiyi Zhang, Eunyee Koh et al.

Figures, such as bar charts, pie charts, and line plots, are widely used to convey important information in a concise format. They are usually human-friendly but difficult for computers to process automatically. In this work, we investigate the problem of figure captioning where the goal is to automatically generate a natural language description of the figure. While natural image captioning has been studied extensively, figure captioning has received relatively little attention and remains a challenging problem. First, we introduce a new dataset for figure captioning, FigCAP, based on FigureQA. Second, we propose two novel attention mechanisms. To achieve accurate generation of labels in figures, we propose Label Maps Attention. To model the relations between figure labels, we propose Relation Maps Attention. Third, we use sequence-level training with reinforcement learning in order to directly optimizes evaluation metrics, which alleviates the exposure bias issue and further improves the models in generating long captions. Extensive experiments show that the proposed method outperforms the baselines, thus demonstrating a significant potential for the automatic captioning of vast repositories of figures.

Charles Chen, Razvan Bunescu

We describe a new semantic parsing setting that allows users to query the system using both natural language questions and actions within a graphical user interface. Multiple time series belonging to an entity of interest are stored in a database and the user interacts with the system to obtain a better understanding of the entity's state and behavior, entailing sequences of actions and questions whose answers may depend on previous factual or navigational interactions. We design an LSTM-based encoder-decoder architecture that models context dependency through copying mechanisms and multiple levels of attention over inputs and previous outputs. When trained to predict tokens using supervised learning, the proposed architecture substantially outperforms standard sequence generation baselines. Training the architecture using policy gradient leads to further improvements in performance, reaching a sequence-level accuracy of 88.7% on artificial data and 74.8% on real data.

Kristen Masada, Razvan Bunescu

We present a new approach to harmonic analysis that is trained to segment music into a sequence of chord spans tagged with chord labels. Formulated as a semi-Markov Conditional Random Field (semi-CRF), this joint segmentation and labeling approach enables the use of a rich set of segment-level features, such as segment purity and chord coverage, that capture the extent to which the events in an entire segment of music are compatible with a candidate chord label. The new chord recognition model is evaluated extensively on three corpora of classical music and a newly created corpus of rock music. Experimental results show that the semi-CRF model performs substantially better than previous approaches when trained on a sufficient number of labeled examples and remains competitive when the amount of training data is limited.

Reza Katebi, Yadi Zhou, Ryan Chornock et al.

Understanding morphological types of galaxies is a key parameter for studying their formation and evolution. Neural networks that have been used previously for galaxy morphology classification have some disadvantages, such as not being invariant under rotation. In this work, we studied the performance of Capsule Network, a recently introduced neural network architecture that is rotationally invariant and spatially aware, on the task of galaxy morphology classification. We designed two evaluation scenarios based on the answers from the question tree in the Galaxy Zoo project. In the first scenario, we used Capsule Network for regression and predicted probabilities for all of the questions. In the second scenario, we chose the answer to the first morphology question that had the highest user agreement as the class of the object and trained a Capsule Network classifier, where we also reconstructed galaxy images. We achieved promising results in both of these scenarios. Automated approaches such as the one introduced here will greatly decrease the workload of astronomers and will play a critical role in the upcoming large sky surveys.

Alexander Bagnall, Razvan Bunescu, Gordon Stewart

We propose a new ensemble method for detecting and classifying adversarial examples generated by state-of-the-art attacks, including DeepFool and C&W. Our method works by training the members of an ensemble to have low classification error on random benign examples while simultaneously minimizing agreement on examples outside the training distribution. We evaluate on both MNIST and CIFAR-10, against oblivious and both white- and black-box adversaries.