Welcome to
Women in AI Research

Can a language model learn a language no human ever could? Stanford PhD researcher Julie Kallini put Noam Chomsky's claim to the test, and won the ACL 2024 Best Paper Award for it.In this episode of Women in AI Research, Julie takes us from "impossible languages" to the hidden costs of tokenization, and shows how byte-level models could make LLMs faster, fairer across languages, and smarter about where they spend compute. We also talk about choosing a PhD advisor, leaving a software engineering job at Meta for a PhD, and what doing research at FAIR is really like.In this episode:Why it would be unethical to raise a child on an invented language, but not a language modelWhy linguists still can't agree on what makes a language "impossible"Why ChatGPT couldn't count the R's in "strawberry", and what tokenization has to do with itHow tokenization can make the same request cost more in some languages than othersMrT5: teaching a byte-level model to decide which bytes it can safely dropFast BLT: speeding up byte-level generation with diffusion and self-speculationDiffusion vs. autoregressive generation: practical advice for practitionersWhat's next: adaptive compression, or why spend dozens of tokens on the MIT license?📄 Papers discussedMission: Impossible Language Models (ACL 2024, Best Paper)MrT5: Dynamic Token Merging for Efficient Byte-level Language Models (ICLR 2025)Fast Byte Latent TransformerByte Latent Transformer: Patches Scale Better Than Tokens (Meta FAIR)Information Locality as an Inductive Bias for Neural Language Models (Someya et al., ACL 2025)👩‍🔬 About Julie KalliniJulie is a fourth-year Computer Science PhD student at Stanford, advised by Chris Potts and Dan Jurafsky. Her research spans computational linguistics, tokenization, efficient language model architectures, and byte-level modeling.🎙️ About WiAIRWomen in AI Research (WiAIR) features in-depth conversations with women shaping AI research: their work, their ideas, and the paths that led them there.🔔 Subscribe for new episodes

Why don't bigger LLMs look more like the human brain? In the brain's language network, today's large models explain only slightly more variance than GPT-2 XL - and the reason says a lot about what those brain regions actually do.Dr. Greta Tuckute (Research Fellow at the Kempner Institute for the Study of Natural and Artificial Intelligence, Harvard) joins Jekaterina Novikova on Women in AI Research to unpack what brain-LLM alignment does and does not tell us. Greta works where neuroscience, cognitive science and AI meet - and when asked to keep only one of those three labels, AI is the first one she drops.The conversation covers how brain alignment develops over training, how sparse autoencoders can turn the "you are just comparing one black box to another" critique into something testable, and what memory experiments reveal about how meaning is stored - from why "pineapple" sticks in memory and "light" does not, to which sentence embeddings predict what people remember.In this episode:Brain alignment tracks formal linguistic competence, not reasoning - and it emerges after not much more than a developmentally plausible ~100M tokens, not 300BWhy better next-word prediction stops meaning "more brain-like" once a model has mastered languageSparse autoencoder features plus surprisal: for one frontal brain region, surprisal alone does almost as well as 32,000 SAE features, while some voxels are captured by just six features- Brains and LLMs share the main, high-variance features of language - not the idiosyncratic onesWhy learning from BPE tokens puts brain-LLM comparisons on "pretty shaky ground", and what should come nextA distinctive meaning makes words and sentences memorable - and SBERT predicts human sentence memory better than the other embedding models testedFrom running a photography business at 14 to a PhD at MIT, and why trying the other path first was worth itPAPERS DISCUSSEDFrom Language to Cognition: How LLMs Outgrow the Human Language NetworkInterpreting Brain Responses to Language with Sparse Features from Language ModelsDriving and suppressing the human language network using large language models Intrinsically memorable words have unique associations with their meaningsA distinctive meaning makes a sentence memorableMENTIONED IN THIS EPISODEAnna Ivanova on formal vs functional linguistic competence (WiAIR)GRETA TUCKUTEWebsite: http://www.tuckute.comBluesky: https://bsky.app/profile/gretatuckute.bsky.socialX: https://x.com/GretaTuckuteWomen in AI Research (WiAIR) is a podcast and YouTube channel where Jekaterina Novikova talks with women doing AI research about their work, the questions driving it, and the paths that brought them there.🎧 Subscribe to stay updated on new episodes spotlighting brilliant women shaping the future of AI.Follow WiAIR at:⁠⁠⁠LinkedIn⁠⁠⁠⁠⁠⁠Bluesky⁠⁠⁠⁠⁠⁠X (Twitter)⁠⁠⁠⁠⁠⁠⁠WiAIR website⁠

An image is not worth a thousand words - it's worth an indefinite number of them. So why do the metrics we use to evaluate AI-generated image descriptions still assume there's one correct answer?In this episode of Women in AI Research, I talk with Elisa Kreiss (Assistant Professor of Communication at UCLA, director of the Coalas Lab) about what happens when you actually test the metrics the field relies on, and why CLIPScore, one of the most widely used measures for scoring image descriptions, stops correlating with human judgment the moment you introduce context. We also get into why longer descriptions aren't necessarily more informative, what happens when you just ask a model to "be concise," and whether AI models trip over charts and graphs the same way humans do.Elisa's research sits at the intersection of linguistics, accessibility, and multimodal AI, and this conversation covers the full arc of her work, from the theoretical question of why humans never describe images the same way twice, to the practical question of what that means for building systems that actually work for blind and low-vision users.In this episode:Why "context matters" is more radical than it sounds for image description evaluationThe hidden reason CLIPScore breaks down once context enters the pictureWhy length is a bad proxy for information density - and what to use insteadWhat happens when you prompt a model to just "be concise"Why charts and photos need completely different evaluation approachesWhether AI models make the same mistakes as humans when reading data visualizationsWhat NeurIPS's Top Reviewer Award taught her about writing a genuinely useful peer reviewResources & Links:Context Matters for Image Descriptions for Accessibility: Challenges for Referenceless Evaluation MetricsWhen More Words Say Less: Decoupling Length and Specificity in Image Description EvaluationCHART-6: Human-Centered Evaluation of Data Visualization Understanding in Vision-Language Models🎧 Subscribe to stay updated on new episodes spotlighting brilliant women shaping the future of AI.Follow WiAIR at:⁠⁠LinkedIn⁠⁠⁠⁠Bluesky⁠⁠⁠⁠X (Twitter)⁠⁠⁠⁠⁠WiAIR website⁠