Biological and artificial information processing systems form representations of the world that they can use to categorize, reason, plan, navigate, and make decisions. To what extent do the representations formed by these diverse systems agree? Can diverging representations still lead to the same behaviors? And how can systems modify their representations to better match those of another system? These questions pertaining to the study of representational alignment are at the heart of some of the most active research areas in contemporary cognitive science, neuroscience, and machine learning. However, despite this shared goal, the machine learning, neuroscience, and cognitive science communities that study alignment currently lack a shared framework for conveying insights across methodologies and disciplines. This workshop aims to bridge this gap by defining, evaluating, and understanding the implications of representational alignment among biological & artificial systems.

Climate change is a complex global challenge with increasingly severe consequences for humanity as natural disasters multiply, sea levels rise, and ecosystems falter. Collective and urgent action is necessary to limit the extent of climate change impacts and adapt to their effects. Such action can take many forms, from designing smart electric grids (Nweye et al. [2023]) to tracking greenhouse gas emissions through satellite imagery (Bonczak et al. [2023]). Machine learning (ML) can be one useful tool for tackling climate change across multiple scales and sectors via mitigation and adaptation (Rolnick et al. [2022]). Success in both strategies requires encouragement of closer collaboration between diverse disciplines and stakeholders. This workshop intends to bring together those applying ML to climate change challenges and facilitate cross-pollination between ML researchers and experts in complementary climate-relevant fields. Building on our past workshops on this topic, we specifically focus on two aspects that fosters the maturity of ML applications for tackling climate change. The workshop will shed light on work that deploys, analyzes or critiques ML methods and their use for climate change adaptation and mitigation. In addition, we will discuss the key investment mechanics and policy frameworks needed for these applications to leapfrog towards …

Over the past decade, the integration of Artificial Intelligence (AI) for scientific exploration has grown as a transformative force, propelling research into new realms of discovery. The AI4DifferentialEquations in Science workshop at ICLR 2024 invites participants on a dynamic journey at the interface of machine learning and computational sciences known as Scientific Machine Learning (SciML).This workshop aims to unleash innovative approaches that harness the power of AI algorithms combined with computational mathematics to advance scientific discovery and problem solving. This enables us to push the boundaries of scientific computing beyond its traditional limits. Our goal is to delve into the latest AI advancements, particularly those that significantly enhance the efficiency of solving ordinary and partial differential equations (PDEs). These methods result in significant performance gains, which allow for solutions at high resolution that were previously unfeasible or required large amounts of computation. The AI4DifferentialEquations in Science workshop aims to unlock the full potential of data-driven approaches in advancing scientific frontiers in earth sciences, climate and computational fluid dynamics to name a few.Key topics include but are not limited to:- Exploration of novel applications of deep learning techniques in scientific simulations of partial or ordinary differential equations.- Forward and inverse problems …

Recent advances in artificial intelligence greatly benefit from data-driven machine learning methods that train deep neural networks with large scale data. The usage of data should be responsible, transparent, and comply with privacy regulations. This workshop aims to bring together industry and academic researchers, privacy regulators and legal, policy people to have a conversation on privacy research. We hope to (re)visit major privacy considerations from both technical and nontechnical perspectives through discussions with interdisciplinary discussions. Topics of interest include, but are not limited toRelationship of privacy regulation (such as GDPR, DMA) to machine learning;Interpolation and explanation of data privacy;Efficient methods for privacy preserving machine learning;Federated learning for data minimization;Differential privacy theory and practice;Threat model and privacy attacks;Encryption methods for machine learning;Privacy in machine learning systems;Privacy for large language models;Relationship between privacy, transparency, auditability, verifiability;Relationship between privacy, robustness, fairness etc.

The constant and breakneck speed of progress being made in artificial intelligence (AI) and generative AI needs to be resource optimized for practical societal impacts. Adapting the state-of-the-art (SOTA) methods such as large language models (LLMs), Diffusion Models, and Neural Radiance Fields (NeRFs) to resource-constrained environments, to run (even few-show fine-tuning and inference) under low resources such as those typically in developing countries and computing at the edge, is highly challenging in practice. Partly due to the lack of diversity in the data and personnels (involved in annotating and validating), their high demand in computational resources, variations in the selection of performance metrics. For example, recent breakthroughs in natural language processing (NLP), computer vision, and speech analysis, for instance, rely on increasingly complex and large models (e.g. most models based on transformers and attention such as BERT, GPT-2/GPT-3, DALLE-2, and stable diffusion) that are pre-trained in on large corpus of unlabeled data. Applying these models in a resource-constrained environment is a non-trivial challenge. Moreover, the potential risks associated with such large models in low resource settings, e.g., disinformation, is virtually unexplored. Low/limited resources mean a hard path towards the adoption of these breakthroughs for most edge applications as well as …

Generative Artificial Intelligence (AI) has made significant advancements in recent years, particularly with the development of large language and diffusion models. These generative models have demonstrated impressive capabilities in various tasks, such as text generation and image and audio synthesis. Concurrently, Reinforcement Learning (RL) has made significant strides in solving complex sequential decision-making problems with the help of external knowledge sources . However, there remains untapped potential in combining generative models with RL algorithms to tackle real-world challenges, particularly to improve sample efficiency of tabula rasa training by introducing priors from related domains such as visual question-answering, image captioning and image generation.This workshop aims to bring together researchers and practitioners from the fields of generative AI and reinforcement learning to explore the latest advances, methodologies, and applications. By fostering collaborations between these two domains, we intend to unlock new opportunities for addressing complex problems that lie at the intersection of both fields.

This workshop delves into the significance of agents driven by large language models (LLMs), a topic that has recently sparked intense discussions. Building on the current huge progress on LLMs, we'll focus on autonomous agents that perform intricate tasks in both real and simulated environments guided by natural language instructions. What sets these agents apart is their sophisticated use of language prompts, not just as a means of communication but also as a medium for reasoning—a characteristic once thought unique to humans. Our workshop specifically aims to discuss the methods, tasks, theories, and risks associated with LLM-driven agents that are capable of using language as a tool for thought and communication.

The dream of Artificial General Intelligence (AGI) has been the north star of AI research since its inception. While large language models like GPT-4 and LLama-2 signal promising advances, the road to true AGI remains complex, mixed with both promise and challenges. Our workshop offers a collaborative platform to discuss AGI's potential, its historical roots, and pressing challenges, blending insights from both academia and industry. Adopting a hybrid format, we invite global researchers to join, share, and shape the discourse on our journey towards AGI.We invite submissions on a range of topics including, but not limited to:- Frontiers of AGI research- Classic AGI Attempts as Inspiration- Interdisciplinary Insights for AGI- Fundamental Limitations of LLMs- Practical Limitations of LLMs- Safety, Ethics, and Regulation in AGI Development- AGI's Economic and Societal Impacts

Models and methods based on large-scale foundation models (FMs) are dominating a large variety of applications in natural language processing, computer vision and other domains. These models, with their immense capabilities, offer a plethora of benefits but also introduce challenges related to reliability, transparency, and ethics. The workshop on reliable and responsible FMs addresses the urgent need to ensure that such models are trustworthy, robust and aligned with human values. The significance of this topic cannot be overstated, as the real-world implications of foundation models impact everything from daily information access to critical decision-making in fields ranging from medicine to finance. The responsible design, deployment, and oversight of these models dictate not only the success of AI solutions but also the preservation of societal norms, equity, and fairness. Moreover, these issues will become increasingly more important in the future, as the capabilities and adoption of FMs increase.

This in-person workshop delves into the intricacies of extracting valuable insights from healthcare time series data, aiming to deepen our comprehension of human health. The session will concentrate on two important themes within this domain: Behavioral Health and Foundational Models. Throughout the workshop, we will highlight cutting-edge approaches and solutions tailored for analyzing healthcare time series data, positioning attendees at the forefront of advancements in this field. By participating in this workshop, attendees will also become part of a diverse community committed to driving forward the boundaries of knowledge in human health.

The success of deep learning practices has driven the rapid development of learning theory. However, recent studies have pointed out that contrasting scenarios and conclusions exist between many existing theories and their corresponding real-world applications, leading to a significant gap. This workshop aims to bridge this gap by (i) troubleshooting unnoticed gaps between learning theory and practice and (ii) narrowing the existing ones by developing new analyses. We hope that this workshop will not only raise awareness of the challenges in bridging the gap between theory and practice in deep learning but also inspire new solutions and insights that contribute to the advancement of deep learning.The workshop website: https://sites.google.com/view/bgpt-iclr24.

This workshop focuses on the increasing role of data in machine learning, particularly in the context of science research and application. It builds on existing momentum in the field and explores important topics including data quality, governance, ethics, infrastructure & tools, and community development. The main objective of the workshop is to shape the agenda for data-centric machine learning research (DMLR). Towards that end, the workshop will foster collaboration among researchers, practitioners, data producers, data consumers, and experts in DMLR. The workshop will feature keynote talks, panel discussions, poster sessions, and networking opportunities.

Promoting diverse viewpoints and trans-disciplinary research is the basis for addressing the pressing questions of our times, such as climate change, social inequalities, biodiversity, and food security. Developing modern machine learning approaches tailored towards remote sensing data is key to investigating these problems efficiently. This second Machine Learning for Remote Sensing (ML4RS) workshop promotes this exchange by allowing researchers to present their a) research on environmentally and societally important applications and/or b) innovative methods that can have an impact in such application domains. This workshop is the continuation of the ICLR 2023 ML for Remote Sensing workshop that enabled local stakeholders, researchers, and students, for instance, from the Rwanda Space Agency and local CMU Africa, to discuss and debate the key problems to be addressed with machine learning. In this workshop in Vienna, we continue in this spirit by giving locally-based experts a voice to articulate important regional challenges in ML4RS, for instance, by inviting domain scientists from international organizations, such as the IAEA or the Red Cross (invited panelists). The keynote speakers are leading researchers in the intersection of machine learning and remote sensing. Our workshop is financially sponsored by industry and governmental organizations, such as the European Space …

Biomolecular design, through artificial engineering of proteins, ligands, and nucleic acids, holds immense promise in addressing pressing medical, industrial, and environmental challenges. While generative machine learning has shown significant potential in this area, a palpable disconnect exists with experimental biology: many ML research efforts prioritize static benchmark performance, potentially sidelining impactful biological applications. This workshop seeks to bridge this gap by bringing computationalists and experimentalists together, catalyzing a deeper interdisciplinary discourse. Together, we will explore the strengths and challenges of generative ML in biology, experimental integration of generative ML, and pinpoint biological problems ready for ML. To attract high-quality and diverse research, we partnered with Cell Systems for a special collection, and we created dedicated tracks for in-silico ML research and hybrid ML-experimental biology research. Our lineup features renowned scientists as speakers and emerging leaders as panelists, encapsulating a spectrum from high-throughput experimentation and computational biology to generative ML. With a diverse organizing team and backed by industry sponsors, we dedicate the workshop to pushing the boundaries of ML's role in biology.

Over 1 billion people live in Africa, and its residents speak more than 2,000 languages. But those languages are among the least represented in NLP research, and work on African languages is often sidelined at major venues. Over the past few years, a vibrant, collaborative community of researchers has formed around a sustained focus on NLP for the benefit of the African continent: national, regional, continental and even global collaborative efforts focused on African languages, African corpora, and tasks with importance in the African context. The AfricaNLP workshops have been a central venue in organizing, sustaining, and growing this focus, and we propose to continue this tradition with an AfricaNLP 2024 workshop in Vienna.Starting in 2020, the AfricaNLP workshop has become a core event for the African NLP community and has drawn global attendance and interest. Many of the participants are active in the Masakhane grassroots NLP community, allowing the community to convene, showcase and share experiences with each other. Large scale collaborative works have been enabled by participants who joined from the AfricaNLP workshop such as MasakhaNER (61 authors), Quality assessment of Multilingual Datasets (51 authors), Corpora Building for Twi (25 authors), NLP for Ghanaian Languages (25 Authors). Many …

Large Language Models (LLMs) have emerged as transformative tools in natural language processing, redefining benchmarks across tasks from machine translation to dialog systems. However, with these advancements come intricate challenges centered around the security, transparency, and ethical dimensions of LLMs. These challenges, ranging from biases and misinformation dissemination to vulnerabilities against sophisticated attacks, have garnered considerable research attention. Our proposed workshop seeks to shine a spotlight on these pivotal issues, focusing on a myriad of topics including, but not limited to, LLM reliability, interpretability, backdoor defenses, and emerging learning paradigms. This assembly aims to bridge gaps between academia and industry, offering a platform for rigorous discussion, collaborative brainstorming, and a showcase of the latest research breakthroughs. Through this endeavor, we aspire to pave a pathway towards more secure, transparent, and ethically-grounded developments in LLMs, underlining the importance of collaborative, cross-disciplinary efforts in the process.

The critical bottleneck in drug discovery is still our limited understanding of the biological mechanisms underlying diseases. Consequently, often we do not know why patients develop specific diseases, and many drug candidates fail in clinical trials. Recent advancements in new genomics platforms and the development of diverse omics datasets have ignited a growing interest in the study of this field. In addition, machine learning plays a pivotal role in improving success rates in language processing, image analysis, and molecular design. The boundaries between these two domains are becoming increasingly blurred, particularly with the emergence of modern foundation models that stand at the intersection of data-driven approaches, self-supervised techniques, and genomic explorations. This workshop aims to elucidate the intricate relationship between genomics, target identification, and fundamental machine learning methods. By strengthening the connection between machine learning and target identification via genomics, new possibilities for interdisciplinary research in these areas will emerge.

Building globally-inclusive generative artificial intelligence (genAI) that encodes, respects, and valorizes cultural sensibilities as well as performs well for users across cultural contexts, is an important goal as we deploy generative AI products globally. If we are to build such inclusive AI for people, we must both have a better understanding of how we can make our AI pipeline more globally inclusive and how AI technologies can impact or shape cultures it is deployed in. If unexamined, this relationship between AI and culture will universalize western-centered AI and have unforeseen impacts on global cultural production, values and consumptions. However, this is not a relationship AI scholarship can understand on its own. We need an engagement with existing theories on the interplay between technology and culture, and how both can shape each other. We urgently thus need a cross-disciplinary and cross-community framework for understanding this multifaceted relationship between AI and Culture. This workshop aims to begin a conversation between core AI researchers and experts from the social sciences and humanities with a focus on the impact of generative AI on cultures and the cultural exclusions embedded in our on generative AI pipelines. Through this focus, the workshop will encourage field building …

Foundation models (FMs) have revolutionized machine learning research across domains. These models are trained on extensive, highly varied datasets and can be quickly adapted to solve many tasks of interest. FMs are extremely effective on language (e.g., GPT-3 [1], BERT [2], PaLM [3], LLaMa [17]), vision (e.g., SimCLR [4]), speech (e.g., Whisper), and multi-modal (e.g., CLIP [5], DALL-E [6]) inputs.However, understanding of FMs lags far behind their extraordinary performance. FMs are known for their surprising emergent capabilities, such as in-context learning [1], but rigorous characterization of such phenomena is sorely lacking. Recently, substantially smaller models (e.g., LLaMA [17]) have demonstrated performance comparable to or better than huge FMs from the previous generation (e.g, OPT [19]). These findings suggest that careful selection of data, training objectives, and adaptation methods can more effectively induce desirable properties in FMs. Development of such techniques can be accelerated through better understanding.This workshop aims to bring together researchers who work on developing an understanding of FMs, through either careful experimentation or theoretical work. Rigorous characterization of FMs can also contribute to the broader goal of mitigating undesirable behaviors. FMs are now broadly available to users, so misaligned models present real-world risk. We thus also welcome submissions …

Foundation Models (FMs, e.g., GPT-3/4, LLaMA, DALL-E, Stable Diffusion, etc.) have been achieving sweeping success on a wide range of tasks. As researchers strive to keep up with the understanding of the capabilities and limitations of FMs as well as their implications following the rapid evolution, the attention is now shifting to the emerging notion of data-centric AI. The curation of training data has been shown to be crucially important for the performance and reliability of FMs and a wealth of recent works demonstrate that data-perspective research sheds light on a promising direction toward critical issues such as safety, alignment, efficiency, security, privacy, interpretability, etc. Recent year has seen a spur of individual works exploring many frontiers related to this topic, providing now an excellent opportunity to bring together brilliant minds to search for a systematic framework and roadmap for research. This workshop aims to discuss and explore a better understanding of the new paradigm for research on data problems for foundation models. Our technical agenda is composed of four modules with 12 confirmed speakers:- A. Data Quality, Dataset Curation, and Data Generation–Recent Achievements and Current Efforts- B. A Data Perspective to Efficiency, Interpretability, and Alignment–Latest Advancement and Breakthroughs- C. …