Hendrik Strobelt

BigScience Workshop, Teven Le Scao, Angela Fan et al. · allen-ai, berkeley

Large language models (LLMs) have been shown to be able to perform new tasks based on a few demonstrations or natural language instructions. While these capabilities have led to widespread adoption, most LLMs are developed by resource-rich organizations and are frequently kept from the public. As a step towards democratizing this powerful technology, we present BLOOM, a 176B-parameter open-access language model designed and built thanks to a collaboration of hundreds of researchers. BLOOM is a decoder-only Transformer language model that was trained on the ROOTS corpus, a dataset comprising hundreds of sources in 46 natural and 13 programming languages (59 in total). We find that BLOOM achieves competitive performance on a wide variety of benchmarks, with stronger results after undergoing multitask prompted finetuning. To facilitate future research and applications using LLMs, we publicly release our models and code under the Responsible AI License.

Hendrik Strobelt, Albert Webson, Victor Sanh et al. · deepmind, ibm-research

State-of-the-art neural language models can now be used to solve ad-hoc language tasks through zero-shot prompting without the need for supervised training. This approach has gained popularity in recent years, and researchers have demonstrated prompts that achieve strong accuracy on specific NLP tasks. However, finding a prompt for new tasks requires experimentation. Different prompt templates with different wording choices lead to significant accuracy differences. PromptIDE allows users to experiment with prompt variations, visualize prompt performance, and iteratively optimize prompts. We developed a workflow that allows users to first focus on model feedback using small data before moving on to a large data regime that allows empirical grounding of promising prompts using quantitative measures of the task. The tool then allows easy deployment of the newly created ad-hoc models. We demonstrate the utility of PromptIDE (demo at http://prompt.vizhub.ai) and our workflow using several real-world use cases.

Sebastian Gehrmann, Abhik Bhattacharjee, Abinaya Mahendiran et al. · amazon-science, cmu

Evaluation in machine learning is usually informed by past choices, for example which datasets or metrics to use. This standardization enables the comparison on equal footing using leaderboards, but the evaluation choices become sub-optimal as better alternatives arise. This problem is especially pertinent in natural language generation which requires ever-improving suites of datasets, metrics, and human evaluation to make definitive claims. To make following best model evaluation practices easier, we introduce GEMv2. The new version of the Generation, Evaluation, and Metrics Benchmark introduces a modular infrastructure for dataset, model, and metric developers to benefit from each others work. GEMv2 supports 40 documented datasets in 51 languages. Models for all datasets can be evaluated online and our interactive data card creation and rendering tools make it easier to add new datasets to the living benchmark.

Benjamin Hoover, Yuchen Liang, Bao Pham et al. · gatech, ibm-research

Our work combines aspects of three promising paradigms in machine learning, namely, attention mechanism, energy-based models, and associative memory. Attention is the power-house driving modern deep learning successes, but it lacks clear theoretical foundations. Energy-based models allow a principled approach to discriminative and generative tasks, but the design of the energy functional is not straightforward. At the same time, Dense Associative Memory models or Modern Hopfield Networks have a well-established theoretical foundation, and allow an intuitive design of the energy function. We propose a novel architecture, called the Energy Transformer (or ET for short), that uses a sequence of attention layers that are purposely designed to minimize a specifically engineered energy function, which is responsible for representing the relationships between the tokens. In this work, we introduce the theoretical foundations of ET, explore its empirical capabilities using the image completion task, and obtain strong quantitative results on the graph anomaly detection and graph classification tasks.

Angie Boggust, Harini Suresh, Hendrik Strobelt et al. · mit

Saliency methods are a common class of machine learning interpretability techniques that calculate how important each input feature is to a model's output. We find that, with the rapid pace of development, users struggle to stay informed of the strengths and limitations of new methods and, thus, choose methods for unprincipled reasons (e.g., popularity). Moreover, despite a corresponding rise in evaluation metrics, existing approaches assume universal desiderata for saliency methods (e.g., faithfulness) that do not account for diverse user needs. In response, we introduce saliency cards: structured documentation of how saliency methods operate and their performance across a battery of evaluative metrics. Through a review of 25 saliency method papers and 33 method evaluations, we identify 10 attributes that users should account for when choosing a method. We group these attributes into three categories that span the process of computing and interpreting saliency: methodology, or how the saliency is calculated; sensitivity, or the relationship between the saliency and the underlying model and data; and, perceptibility, or how an end user ultimately interprets the result. By collating this information, saliency cards allow users to more holistically assess and compare the implications of different methods. Through nine semi-structured interviews with users from various backgrounds, including researchers, radiologists, and computational biologists, we find that saliency cards provide a detailed vocabulary for discussing individual methods and allow for a more systematic selection of task-appropriate methods. Moreover, with saliency cards, we are able to analyze the research landscape in a more structured fashion to identify opportunities for new methods and evaluation metrics for unmet user needs.

Furui Cheng, Vilém Zouhar, Simran Arora et al. · eth-zurich

Large Language Models (LLMs) are notorious for blending fact with fiction and generating non-factual content, known as hallucinations. To address this challenge, we propose an interactive system that helps users gain insight into the reliability of the generated text. Our approach is based on the idea that the self-consistency of multiple samples generated by the same LLM relates to its confidence in individual claims in the generated texts. Using this idea, we design RELIC, an interactive system that enables users to investigate and verify semantic-level variations in multiple long-form responses. This allows users to recognize potentially inaccurate information in the generated text and make necessary corrections. From a user study with ten participants, we demonstrate that our approach helps users better verify the reliability of the generated text. We further summarize the design implications and lessons learned from this research for future studies of reliable human-LLM interactions.

Benjamin Hoover, Hendrik Strobelt, Dmitry Krotov et al. · gatech, ibm-research

The generative process of Diffusion Models (DMs) has recently set state-of-the-art on many AI generation benchmarks. Though the generative process is traditionally understood as an "iterative denoiser", there is no universally accepted language to describe it. We introduce a novel perspective to describe DMs using the mathematical language of memory retrieval from the field of energy-based Associative Memories (AMs), making efforts to keep our presentation approachable to newcomers to both of these fields. Unifying these two fields provides insight that DMs can be seen as a particular kind of AM where Lyapunov stability guarantees are bypassed by intelligently engineering the dynamics (i.e., the noise and step size schedules) of the denoising process. Finally, we present a growing body of evidence that records DMs exhibiting empirical behavior we would expect from AMs, and conclude by discussing research opportunities that are revealed by understanding DMs as a form of energy-based memory.

Richard Brath, Daniel Keim, Johannes Knittel et al.

With a constant increase of learned parameters, modern neural language models become increasingly more powerful. Yet, explaining these complex model's behavior remains a widely unsolved problem. In this paper, we discuss the role interactive visualization can play in explaining NLP models (XNLP). We motivate the use of visualization in relation to target users and common NLP pipelines. We also present several use cases to provide concrete examples on XNLP with visualization. Finally, we point out an extensive list of research opportunities in this field.

Dennis Wei, Ronny Luss, Xiaomeng Hu et al.

Large Language Models (LLMs) have become ubiquitous in everyday life and are entering higher-stakes applications ranging from summarizing meeting transcripts to answering doctors' questions. As was the case with earlier predictive models, it is crucial that we develop tools for explaining the output of LLMs, be it a summary, list, response to a question, etc. With these needs in mind, we introduce In-Context Explainability 360 (ICX360), an open-source Python toolkit for explaining LLMs with a focus on the user-provided context (or prompts in general) that are fed to the LLMs. ICX360 contains implementations for three recent tools that explain LLMs using both black-box and white-box methods (via perturbations and gradients respectively). The toolkit, available at https://github.com/IBM/ICX360, contains quick-start guidance materials as well as detailed tutorials covering use cases such as retrieval augmented generation, natural language generation, and jailbreaking.

Angie Boggust, Hyemin Bang, Hendrik Strobelt et al. · mit

While interpretability methods identify a model's learned concepts, they overlook the relationships between concepts that make up its abstractions and inform its ability to generalize to new data. To assess whether models' have learned human-aligned abstractions, we introduce abstraction alignment, a methodology to compare model behavior against formal human knowledge. Abstraction alignment externalizes domain-specific human knowledge as an abstraction graph, a set of pertinent concepts spanning levels of abstraction. Using the abstraction graph as a ground truth, abstraction alignment measures the alignment of a model's behavior by determining how much of its uncertainty is accounted for by the human abstractions. By aggregating abstraction alignment across entire datasets, users can test alignment hypotheses, such as which human concepts the model has learned and where misalignments recur. In evaluations with experts, abstraction alignment differentiates seemingly similar errors, improves the verbosity of existing model-quality metrics, and uncovers improvements to current human abstractions.

Walid Bousselham, Angie Boggust, Sofian Chaybouti et al. · mit

Vision Transformers (ViTs), with their ability to model long-range dependencies through self-attention mechanisms, have become a standard architecture in computer vision. However, the interpretability of these models remains a challenge. To address this, we propose LeGrad, an explainability method specifically designed for ViTs. LeGrad computes the gradient with respect to the attention maps of ViT layers, considering the gradient itself as the explainability signal. We aggregate the signal over all layers, combining the activations of the last as well as intermediate tokens to produce the merged explainability map. This makes LeGrad a conceptually simple and an easy-to-implement tool for enhancing the transparency of ViTs. We evaluate LeGrad in challenging segmentation, perturbation, and open-vocabulary settings, showcasing its versatility compared to other SotA explainability methods demonstrating its superior spatial fidelity and robustness to perturbations. A demo and the code is available at https://github.com/WalBouss/LeGrad.

Lucas Monteiro Paes, Dennis Wei, Hyo Jin Do et al. · harvard

Despite the increasing use of large language models (LLMs) for context-grounded tasks like summarization and question-answering, understanding what makes an LLM produce a certain response is challenging. We propose Multi-Level Explanations for Generative Language Models (MExGen), a technique to provide explanations for context-grounded text generation. MExGen assigns scores to parts of the context to quantify their influence on the model's output. It extends attribution methods like LIME and SHAP to LLMs used in context-grounded tasks where (1) inference cost is high, (2) input text is long, and (3) the output is text. We conduct a systematic evaluation, both automated and human, of perturbation-based attribution methods for summarization and question answering. The results show that our framework can provide more faithful explanations of generated output than available alternatives, including LLM self-explanations. We open-source code for MExGen as part of the ICX360 toolkit: https://github$.$com/IBM/ICX360.

Seongmin Lee, Benjamin Hoover, Hendrik Strobelt et al. · gatech, ibm-research

Diffusion-based generative models' impressive ability to create convincing images has garnered global attention. However, their complex internal structures and operations often pose challenges for non-experts to grasp. We introduce Diffusion Explainer, the first interactive visualization tool designed to elucidate how Stable Diffusion transforms text prompts into images. It tightly integrates a visual overview of Stable Diffusion's complex components with detailed explanations of their underlying operations. This integration enables users to fluidly transition between multiple levels of abstraction through animations and interactive elements. Offering real-time hands-on experience, Diffusion Explainer allows users to adjust Stable Diffusion's hyperparameters and prompts without the need for installation or specialized hardware. Accessible via users' web browsers, Diffusion Explainer is making significant strides in democratizing AI education, fostering broader public access. More than 7,200 users spanning 113 countries have used our open-sourced tool at https://poloclub.github.io/diffusion-explainer/. A video demo is available at https://youtu.be/MbkIADZjPnA.

Alexander M. Rush, Hendrik Strobelt

MiniConf is a framework for hosting virtual academic conferences motivated by the sudden inability for these events to be hosted globally. The framework is designed to be global and asynchronous, interactive, and to promote browsing and discovery. We developed the system to be sustainable and maintainable, in particular ensuring that it is open-source, easy to setup, and scalable on minimal hardware. In this technical report, we discuss design decisions, provide technical detail, and show examples of a case study deployment.

Sebastian Gehrmann, Hendrik Strobelt, Alexander M. Rush

The rapid improvement of language models has raised the specter of abuse of text generation systems. This progress motivates the development of simple methods for detecting generated text that can be used by and explained to non-experts. We develop GLTR, a tool to support humans in detecting whether a text was generated by a model. GLTR applies a suite of baseline statistical methods that can detect generation artifacts across common sampling schemes. In a human-subjects study, we show that the annotation scheme provided by GLTR improves the human detection-rate of fake text from 54% to 72% without any prior training. GLTR is open-source and publicly deployed, and has already been widely used to detect generated outputs

David Bau, Jun-Yan Zhu, Hendrik Strobelt et al.

Generative Adversarial Networks (GANs) have achieved impressive results for many real-world applications. As an active research topic, many GAN variants have emerged with improvements in sample quality and training stability. However, visualization and understanding of GANs is largely missing. How does a GAN represent our visual world internally? What causes the artifacts in GAN results? How do architectural choices affect GAN learning? Answering such questions could enable us to develop new insights and better models. In this work, we present an analytic framework to visualize and understand GANs at the unit-, object-, and scene-level. We first identify a group of interpretable units that are closely related to concepts with a segmentation-based network dissection method. We quantify the causal effect of interpretable units by measuring the ability of interventions to control objects in the output. Finally, we examine the contextual relationship between these units and their surrounding by inserting the discovered concepts into new images. We show several practical applications enabled by our framework, from comparing internal representations across different layers, models, and datasets, to improving GANs by locating and removing artifact-causing units, to interactively manipulating objects in the scene. We will open source our interactive tools to help researchers and practitioners better understand their models.

David Bau, Jun-Yan Zhu, Hendrik Strobelt et al.

Generative Adversarial Networks (GANs) have recently achieved impressive results for many real-world applications, and many GAN variants have emerged with improvements in sample quality and training stability. However, they have not been well visualized or understood. How does a GAN represent our visual world internally? What causes the artifacts in GAN results? How do architectural choices affect GAN learning? Answering such questions could enable us to develop new insights and better models. In this work, we present an analytic framework to visualize and understand GANs at the unit-, object-, and scene-level. We first identify a group of interpretable units that are closely related to object concepts using a segmentation-based network dissection method. Then, we quantify the causal effect of interpretable units by measuring the ability of interventions to control objects in the output. We examine the contextual relationship between these units and their surroundings by inserting the discovered object concepts into new images. We show several practical applications enabled by our framework, from comparing internal representations across different layers, models, and datasets, to improving GANs by locating and removing artifact-causing units, to interactively manipulating objects in a scene. We provide open source interpretation tools to help researchers and practitioners better understand their GAN models.

Swapnaja Achintalwar, Adriana Alvarado Garcia, Ateret Anaby-Tavor et al. · ibm-research

Large language models (LLMs) are susceptible to a variety of risks, from non-faithful output to biased and toxic generations. Due to several limiting factors surrounding LLMs (training cost, API access, data availability, etc.), it may not always be feasible to impose direct safety constraints on a deployed model. Therefore, an efficient and reliable alternative is required. To this end, we present our ongoing efforts to create and deploy a library of detectors: compact and easy-to-build classification models that provide labels for various harms. In addition to the detectors themselves, we discuss a wide range of uses for these detector models - from acting as guardrails to enabling effective AI governance. We also deep dive into inherent challenges in their development and discuss future work aimed at making the detectors more reliable and broadening their scope.

Benjamin Hoover, Duen Horng Chau, Hendrik Strobelt et al. · gatech, ibm-research

Dense Associative Memories are high storage capacity variants of the Hopfield networks that are capable of storing a large number of memory patterns in the weights of the network of a given size. Their common formulations typically require storing each pattern in a separate set of synaptic weights, which leads to the increase of the number of synaptic weights when new patterns are introduced. In this work we propose an alternative formulation of this class of models using random features, commonly used in kernel methods. In this formulation the number of network's parameters remains fixed. At the same time, new memories can be added to the network by modifying existing weights. We show that this novel network closely approximates the energy function and dynamics of conventional Dense Associative Memories and shares their desirable computational properties.

Furui Cheng, Vilém Zouhar, Robin Shing Moon Chan et al. · eth-zurich

Understanding the behavior of large language models (LLMs) is crucial for ensuring their safe and reliable use. However, existing explainable AI (XAI) methods for LLMs primarily rely on word-level explanations, which are often computationally inefficient and misaligned with human reasoning processes. Moreover, these methods often treat explanation as a one-time output, overlooking its inherently interactive and iterative nature. In this paper, we present LLM Analyzer, an interactive visualization system that addresses these limitations by enabling intuitive and efficient exploration of LLM behaviors through counterfactual analysis. Our system features a novel algorithm that generates fluent and semantically meaningful counterfactuals via targeted removal and replacement operations at user-defined levels of granularity. These counterfactuals are used to compute feature attribution scores, which are then integrated with concrete examples in a table-based visualization, supporting dynamic analysis of model behavior. A user study with LLM practitioners and interviews with experts demonstrate the system's usability and effectiveness, emphasizing the importance of involving humans in the explanation process as active participants rather than passive recipients.

Vidya Prasad, Chen Zhu-Tian, Anna Vilanova et al.

Diffusion models have garnered significant attention since they can effectively learn complex multivariate Gaussian distributions, resulting in diverse, high-quality outcomes. They introduce Gaussian noise into training data and reconstruct the original data iteratively. Central to this iterative process is a single Unet, adapting across time steps to facilitate generation. Recent work revealed the presence of composition and denoising phases in this generation process, raising questions about the Unets' varying roles. Our study dives into the dynamic behavior of Unets within denoising diffusion probabilistic models (DDPM), focusing on (de)convolutional blocks and skip connections across time steps. We propose an analytical method to systematically assess the impact of time steps and core Unet components on the final output. This method eliminates components to study causal relations and investigate their influence on output changes. The main purpose is to understand the temporal dynamics and identify potential shortcuts during inference. Our findings provide valuable insights into the various generation phases during inference and shed light on the Unets' usage patterns across these phases. Leveraging these insights, we identify redundancies in GLIDE (an improved DDPM) and improve inference time by ~27% with minimal degradation in output quality. Our ultimate goal is to guide more informed optimization strategies for inference and influence new model designs.

Johannes Knittel, Tushaar Gangavarapu, Hendrik Strobelt et al. · amazon-science

Understanding the general priniciples behind transformer models remains a complex endeavor. Experiments with probing and disentangling features using sparse autoencoders (SAE) suggest that these models might manage linear features embedded as directions in the residual stream. This paper explores the resemblance between decoder-only transformer architecture and vector symbolic architectures (VSA) and presents experiments indicating that GPT-2 uses mechanisms involving nearly orthogonal vector bundling and binding operations similar to VSA for computation and communication between layers. It further shows that these principles help explain a significant portion of the actual neural weights.

Walid Bousselham, Angie Boggust, Hendrik Strobelt et al.

As Vision-Language Models (VLMs) become increasingly sophisticated and widely used, it becomes more and more crucial to understand their decision-making process. Traditional explainability methods, designed for classification tasks, struggle with modern autoregressive VLMs due to their complex token-by-token generation process and intricate interactions between visual and textual modalities. We present DEX-AR (Dynamic Explainability for AutoRegressive models), a novel explainability method designed to address these challenges by generating both per-token and sequence-level 2D heatmaps highlighting image regions crucial for the model's textual responses. The proposed method offers to interpret autoregressive VLMs-including varying importance of layers and generated tokens-by computing layer-wise gradients with respect to attention maps during the token-by-token generation process. DEX-AR introduces two key innovations: a dynamic head filtering mechanism that identifies attention heads focused on visual information, and a sequence-level filtering approach that aggregates per-token explanations while distinguishing between visually-grounded and purely linguistic tokens. Our evaluation on ImageNet, VQAv2, and PascalVOC, shows a consistent improvement in both perturbation-based metrics, using a novel normalized perplexity measure, as well as segmentation-based metrics.

Csaba Both, Benjamin Hoover, Hendrik Strobelt et al.

The ability of a model to learn a task depends strongly on both the task difficulty and the model size. We aim to understand how task difficulty relates to the minimum number of parameters required for learning specific tasks in small transformer models. Our study focuses on the ListOps dataset, which consists of nested mathematical operations. We gradually increase task difficulty by introducing new operations or combinations of operations into the training data. We observe that sum modulo n is the hardest to learn. Curiously, when combined with other operations such as maximum and median, the sum operation becomes easier to learn and requires fewer parameters. We show that joint training not only improves performance but also leads to qualitatively different model behavior. We show evidence that models trained only on SUM might be memorizing and fail to capture the number structure in the embeddings. In contrast, models trained on a mixture of SUM and other operations exhibit number-like representations in the embedding space, and a strong ability to distinguish parity. Furthermore, the SUM-only model relies more heavily on its feedforward layers, while the jointly trained model activates the attention mechanism more. Finally, we show that learning pure SUM can be induced in models below the learning threshold of pure SUM, by pretraining them on MAX+MED. Our findings indicate that emergent abilities in language models depend not only on model size, but also the training curriculum.

Seongmin Lee, Benjamin Hoover, Hendrik Strobelt et al.

Diffusion-based generative models' impressive ability to create convincing images has garnered global attention. However, their complex structures and operations often pose challenges for non-experts to grasp. We present Diffusion Explainer, the first interactive visualization tool that explains how Stable Diffusion transforms text prompts into images. Diffusion Explainer tightly integrates a visual overview of Stable Diffusion's complex structure with explanations of the underlying operations. By comparing image generation of prompt variants, users can discover the impact of keyword changes on image generation. A 56-participant user study demonstrates that Diffusion Explainer offers substantial learning benefits to non-experts. Our tool has been used by over 10,300 users from 124 countries at https://poloclub.github.io/diffusion-explainer/.

Hendrik Strobelt, Benjamin Hoover, Arvind Satyanarayan et al.

While different language models are ubiquitous in NLP, it is hard to contrast their outputs and identify which contexts one can handle better than the other. To address this question, we introduce LMdiff, a tool that visually compares probability distributions of two models that differ, e.g., through finetuning, distillation, or simply training with different parameter sizes. LMdiff allows the generation of hypotheses about model behavior by investigating text instances token by token and further assists in choosing these interesting text instances by identifying the most interesting phrases from large corpora. We showcase the applicability of LMdiff for hypothesis generation across multiple case studies. A demo is available at http://lmdiff.net .

Hendrik Strobelt, Jambay Kinley, Robert Krueger et al.

Table2Text systems generate textual output based on structured data utilizing machine learning. These systems are essential for fluent natural language interfaces in tools such as virtual assistants; however, left to generate freely these ML systems often produce misleading or unexpected outputs. GenNI (Generation Negotiation Interface) is an interactive visual system for high-level human-AI collaboration in producing descriptive text. The tool utilizes a deep learning model designed with explicit control states. These controls allow users to globally constrain model generations, without sacrificing the representation power of the deep learning models. The visual interface makes it possible for users to interact with AI systems following a Refine-Forecast paradigm to ensure that the generation system acts in a manner human users find suitable. We report multiple use cases on two experiments that improve over uncontrolled generation approaches, while at the same time providing fine-grained control. A demo and source code are available at https://genni.vizhub.ai .

Angie Boggust, Benjamin Hoover, Arvind Satyanarayan et al.

Saliency methods -- techniques to identify the importance of input features on a model's output -- are a common step in understanding neural network behavior. However, interpreting saliency requires tedious manual inspection to identify and aggregate patterns in model behavior, resulting in ad hoc or cherry-picked analysis. To address these concerns, we present Shared Interest: metrics for comparing model reasoning (via saliency) to human reasoning (via ground truth annotations). By providing quantitative descriptors, Shared Interest enables ranking, sorting, and aggregating inputs, thereby facilitating large-scale systematic analysis of model behavior. We use Shared Interest to identify eight recurring patterns in model behavior, such as cases where contextual features or a subset of ground truth features are most important to the model. Working with representative real-world users, we show how Shared Interest can be used to decide if a model is trustworthy, uncover issues missed in manual analyses, and enable interactive probing.

Eden Bensaid, Mauro Martino, Benjamin Hoover et al.

Storytelling is an open-ended task that entails creative thinking and requires a constant flow of ideas. Natural language generation (NLG) for storytelling is especially challenging because it requires the generated text to follow an overall theme while remaining creative and diverse to engage the reader. In this work, we introduce a system and a web-based demo, FairyTailor, for human-in-the-loop visual story co-creation. Users can create a cohesive children's fairytale by weaving generated texts and retrieved images with their input. FairyTailor adds another modality and modifies the text generation process to produce a coherent and creative sequence of text and images. To our knowledge, this is the first dynamic tool for multimodal story generation that allows interactive co-formation of both texts and images. It allows users to give feedback on co-created stories and share their results.

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.

Sarah Schwettmann, Hendrik Strobelt, Mauro Martino

In Marius von Senden's Space and Sight, a newly sighted blind patient describes the experience of a corner as lemon-like, because corners "prick" sight like lemons prick the tongue. Prickliness, here, is a dimension in the feature space of sensory experience, an effect of the perceived on the perceiver that arises where the two interact. In the account of the newly sighted, an effect familiar from one interaction translates to a novel context. Perception serves as the vehicle for generalization, in that an effect shared across different experiences produces a concrete abstraction grounded in those experiences. Cezanne and the post-impressionists, fluent in the language of experience translation, realized that the way to paint a concrete form that best reflected reality was to paint not what they saw, but what it was like to see. We envision a future of creation using AI where what it is like to see is replicable, transferrable, manipulable - part of the artist's palette that is both grounded in a particular context, and generalizable beyond it. An active line of research maps human-interpretable features onto directions in GAN latent space. Supervised and self-supervised approaches that search for anticipated directions or use off-the-shelf classifiers to drive image manipulation in embedding space are limited in the variety of features they can uncover. Unsupervised approaches that discover useful new directions show that the space of perceptually meaningful directions is nowhere close to being fully mapped. As this space is broad and full of creative potential, we want tools for direction discovery that capture the richness and generalizability of human perception. Our approach puts creators in the discovery loop during real-time tool use, in order to identify directions that are perceptually meaningful to them, and generate interpretable image translations along those directions.

David Bau, Jun-Yan Zhu, Hendrik Strobelt et al.

Deep neural networks excel at finding hierarchical representations that solve complex tasks over large data sets. How can we humans understand these learned representations? In this work, we present network dissection, an analytic framework to systematically identify the semantics of individual hidden units within image classification and image generation networks. First, we analyze a convolutional neural network (CNN) trained on scene classification and discover units that match a diverse set of object concepts. We find evidence that the network has learned many object classes that play crucial roles in classifying scene classes. Second, we use a similar analytic method to analyze a generative adversarial network (GAN) model trained to generate scenes. By analyzing changes made when small sets of units are activated or deactivated, we find that objects can be added and removed from the output scenes while adapting to the context. Finally, we apply our analytic framework to understanding adversarial attacks and to semantic image editing.

Payel Das, Tom Sercu, Kahini Wadhawan et al.

De novo therapeutic design is challenged by a vast chemical repertoire and multiple constraints, e.g., high broad-spectrum potency and low toxicity. We propose CLaSS (Controlled Latent attribute Space Sampling) - an efficient computational method for attribute-controlled generation of molecules, which leverages guidance from classifiers trained on an informative latent space of molecules modeled using a deep generative autoencoder. We screen the generated molecules for additional key attributes by using deep learning classifiers in conjunction with novel features derived from atomistic simulations. The proposed approach is demonstrated for designing non-toxic antimicrobial peptides (AMPs) with strong broad-spectrum potency, which are emerging drug candidates for tackling antibiotic resistance. Synthesis and testing of only twenty designed sequences identified two novel and minimalist AMPs with high potency against diverse Gram-positive and Gram-negative pathogens, including one multidrug-resistant and one antibiotic-resistant K. pneumoniae, via membrane pore formation. Both antimicrobials exhibit low in vitro and in vivo toxicity and mitigate the onset of drug resistance. The proposed approach thus presents a viable path for faster and efficient discovery of potent and selective broad-spectrum antimicrobials.

David Bau, Hendrik Strobelt, William Peebles et al.

Despite the recent success of GANs in synthesizing images conditioned on inputs such as a user sketch, text, or semantic labels, manipulating the high-level attributes of an existing natural photograph with GANs is challenging for two reasons. First, it is hard for GANs to precisely reproduce an input image. Second, after manipulation, the newly synthesized pixels often do not fit the original image. In this paper, we address these issues by adapting the image prior learned by GANs to image statistics of an individual image. Our method can accurately reconstruct the input image and synthesize new content, consistent with the appearance of the input image. We demonstrate our interactive system on several semantic image editing tasks, including synthesizing new objects consistent with background, removing unwanted objects, and changing the appearance of an object. Quantitative and qualitative comparisons against several existing methods demonstrate the effectiveness of our method.

Vijil Chenthamarakshan, Payel Das, Samuel C. Hoffman et al.

The novel nature of SARS-CoV-2 calls for the development of efficient de novo drug design approaches. In this study, we propose an end-to-end framework, named CogMol (Controlled Generation of Molecules), for designing new drug-like small molecules targeting novel viral proteins with high affinity and off-target selectivity. CogMol combines adaptive pre-training of a molecular SMILES Variational Autoencoder (VAE) and an efficient multi-attribute controlled sampling scheme that uses guidance from attribute predictors trained on latent features. To generate novel and optimal drug-like molecules for unseen viral targets, CogMol leverages a protein-molecule binding affinity predictor that is trained using SMILES VAE embeddings and protein sequence embeddings learned unsupervised from a large corpus. CogMol framework is applied to three SARS-CoV-2 target proteins: main protease, receptor-binding domain of the spike protein, and non-structural protein 9 replicase. The generated candidates are novel at both molecular and chemical scaffold levels when compared to the training data. CogMol also includes insilico screening for assessing toxicity of parent molecules and their metabolites with a multi-task toxicity classifier, synthetic feasibility with a chemical retrosynthesis predictor, and target structure binding with docking simulations. Docking reveals favorable binding of generated molecules to the target protein structure, where 87-95 % of high affinity molecules showed docking free energy < -6 kcal/mol. When compared to approved drugs, the majority of designed compounds show low parent molecule and metabolite toxicity and high synthetic feasibility. In summary, CogMol handles multi-constraint design of synthesizable, low-toxic, drug-like molecules with high target specificity and selectivity, and does not need target-dependent fine-tuning of the framework or target structure information.

David Bau, Jun-Yan Zhu, Jonas Wulff et al.

Despite the success of Generative Adversarial Networks (GANs), mode collapse remains a serious issue during GAN training. To date, little work has focused on understanding and quantifying which modes have been dropped by a model. In this work, we visualize mode collapse at both the distribution level and the instance level. First, we deploy a semantic segmentation network to compare the distribution of segmented objects in the generated images with the target distribution in the training set. Differences in statistics reveal object classes that are omitted by a GAN. Second, given the identified omitted object classes, we visualize the GAN's omissions directly. In particular, we compare specific differences between individual photos and their approximate inversions by a GAN. To this end, we relax the problem of inversion and solve the tractable problem of inverting a GAN layer instead of the entire generator. Finally, we use this framework to analyze several recent GANs trained on multiple datasets and identify their typical failure cases.

Benjamin Hoover, Hendrik Strobelt, Sebastian Gehrmann

Large language models can produce powerful contextual representations that lead to improvements across many NLP tasks. Since these models are typically guided by a sequence of learned self attention mechanisms and may comprise undesired inductive biases, it is paramount to be able to explore what the attention has learned. While static analyses of these models lead to targeted insights, interactive tools are more dynamic and can help humans better gain an intuition for the model-internal reasoning process. We present exBERT, an interactive tool named after the popular BERT language model, that provides insights into the meaning of the contextual representations by matching a human-specified input to similar contexts in a large annotated dataset. By aggregating the annotations of the matching similar contexts, exBERT helps intuitively explain what each attention-head has learned.

Andreas Hinterreiter, Peter Ruch, Holger Stitz et al.

Classifiers are among the most widely used supervised machine learning algorithms. Many classification models exist, and choosing the right one for a given task is difficult. During model selection and debugging, data scientists need to assess classifiers' performances, evaluate their learning behavior over time, and compare different models. Typically, this analysis is based on single-number performance measures such as accuracy. A more detailed evaluation of classifiers is possible by inspecting class errors. The confusion matrix is an established way for visualizing these class errors, but it was not designed with temporal or comparative analysis in mind. More generally, established performance analysis systems do not allow a combined temporal and comparative analysis of class-level information. To address this issue, we propose ConfusionFlow, an interactive, comparative visualization tool that combines the benefits of class confusion matrices with the visualization of performance characteristics over time. ConfusionFlow is model-agnostic and can be used to compare performances for different model types, model architectures, and/or training and test datasets. We demonstrate the usefulness of ConfusionFlow in a case study on instance selection strategies in active learning. We further assess the scalability of ConfusionFlow and present a use case in the context of neural network pruning.

Dylan Cashman, Adam Perer, Remco Chang et al.

Deep learning models require the configuration of many layers and parameters in order to get good results. However, there are currently few systematic guidelines for how to configure a successful model. This means model builders often have to experiment with different configurations by manually programming different architectures (which is tedious and time consuming) or rely on purely automated approaches to generate and train the architectures (which is expensive). In this paper, we present Rapid Exploration of Model Architectures and Parameters, or REMAP, a visual analytics tool that allows a model builder to discover a deep learning model quickly via exploration and rapid experimentation of neural network architectures. In REMAP, the user explores the large and complex parameter space for neural network architectures using a combination of global inspection and local experimentation. Through a visual overview of a set of models, the user identifies interesting clusters of architectures. Based on their findings, the user can run ablation and variation experiments to identify the effects of adding, removing, or replacing layers in a given architecture and generate new models accordingly. They can also handcraft new models using a simple graphical interface. As a result, a model builder can build deep learning models quickly, efficiently, and without manual programming. We inform the design of REMAP through a design study with four deep learning model builders. Through a use case, we demonstrate that REMAP allows users to discover performant neural network architectures efficiently using visual exploration and user-defined semi-automated searches through the model space.

Sebastian Gehrmann, Hendrik Strobelt, Robert Krüger et al.

Automation of tasks can have critical consequences when humans lose agency over decision processes. Deep learning models are particularly susceptible since current black-box approaches lack explainable reasoning. We argue that both the visual interface and model structure of deep learning systems need to take into account interaction design. We propose a framework of collaborative semantic inference (CSI) for the co-design of interactions and models to enable visual collaboration between humans and algorithms. The approach exposes the intermediate reasoning process of models which allows semantic interactions with the visual metaphors of a problem, which means that a user can both understand and control parts of the model reasoning process. We demonstrate the feasibility of CSI with a co-designed case study of a document summarization system.

Atin Sood, Benjamin Elder, Benjamin Herta et al.

Application of neural networks to a vast variety of practical applications is transforming the way AI is applied in practice. Pre-trained neural network models available through APIs or capability to custom train pre-built neural network architectures with customer data has made the consumption of AI by developers much simpler and resulted in broad adoption of these complex AI models. While prebuilt network models exist for certain scenarios, to try and meet the constraints that are unique to each application, AI teams need to think about developing custom neural network architectures that can meet the tradeoff between accuracy and memory footprint to achieve the tight constraints of their unique use-cases. However, only a small proportion of data science teams have the skills and experience needed to create a neural network from scratch, and the demand far exceeds the supply. In this paper, we present NeuNetS : An automated Neural Network Synthesis engine for custom neural network design that is available as part of IBM's AI OpenScale's product. NeuNetS is available for both Text and Image domains and can build neural networks for specific tasks in a fraction of the time it takes today with human effort, and with accuracy similar to that of human-designed AI models.

Cagatay Turkay, Nicola Pezzotti, Carsten Binnig et al.

Data science requires time-consuming iterative manual activities. In particular, activities such as data selection, preprocessing, transformation, and mining, highly depend on iterative trial-and-error processes that could be sped-up significantly by providing quick feedback on the impact of changes. The idea of progressive data science is to compute the results of changes in a progressive manner, returning a first approximation of results quickly and allow iterative refinements until converging to a final result. Enabling the user to interact with the intermediate results allows an early detection of erroneous or suboptimal choices, the guided definition of modifications to the pipeline and their quick assessment. In this paper, we discuss the progressiveness challenges arising in different steps of the data science pipeline. We describe how changes in each step of the pipeline impact the subsequent steps and outline why progressive data science will help to make the process more effective. Computing progressive approximations of outcomes resulting from changes creates numerous research challenges, especially if the changes are made in the early steps of the pipeline. We discuss these challenges and outline first steps towards progressiveness, which, we argue, will ultimately help to significantly speed-up the overall data science process.

Hendrik Strobelt, Sebastian Gehrmann, Michael Behrisch et al.

Neural Sequence-to-Sequence models have proven to be accurate and robust for many sequence prediction tasks, and have become the standard approach for automatic translation of text. The models work in a five stage blackbox process that involves encoding a source sequence to a vector space and then decoding out to a new target sequence. This process is now standard, but like many deep learning methods remains quite difficult to understand or debug. In this work, we present a visual analysis tool that allows interaction with a trained sequence-to-sequence model through each stage of the translation process. The aim is to identify which patterns have been learned and to detect model errors. We demonstrate the utility of our tool through several real-world large-scale sequence-to-sequence use cases.

Hendrik Strobelt, Sebastian Gehrmann, Hanspeter Pfister et al.

Recurrent neural networks, and in particular long short-term memory (LSTM) networks, are a remarkably effective tool for sequence modeling that learn a dense black-box hidden representation of their sequential input. Researchers interested in better understanding these models have studied the changes in hidden state representations over time and noticed some interpretable patterns but also significant noise. In this work, we present LSTMVIS, a visual analysis tool for recurrent neural networks with a focus on understanding these hidden state dynamics. The tool allows users to select a hypothesis input range to focus on local state changes, to match these states changes to similar patterns in a large data set, and to align these results with structural annotations from their domain. We show several use cases of the tool for analyzing specific hidden state properties on dataset containing nesting, phrase structure, and chord progressions, and demonstrate how the tool can be used to isolate patterns for further statistical analysis. We characterize the domain, the different stakeholders, and their goals and tasks.