Videos

See AI2's full collection of videos on our YouTube channel.
Viewing 221-230 of 250 videos
  • Contextual LSTMs A step towards Hierarchial Language Modeling Thumbnail

    Contextual LSTMs A step towards Hierarchial Language Modeling

    September 10, 2015  |  Shalini Ghosh
    Documents exhibit sequential structure at multiple levels of abstraction (e.g., sentences, paragraphs, sections). These abstractions constitute a natural hierarchy for representing the context in which to infer the meaning of words and larger fragments of text. In this talk, we present CLSTM (Contextual LSTM), an…
  • Unsupervised Alignment of Natural Language with Video Thumbnail

    Unsupervised Alignment of Natural Language with Video

    August 18, 2015  |  Iftekhar Naim
    Today we encounter enormous amounts of video data, often accompanied with text descriptions (e.g., cooking videos and recipes, movies and shooting scripts). Extracting meaningful information from these multimodal sequences requires aligning the video frames with the corresponding text sentences. We address the…
  • Feature Generation from Knowledge Graphs Thumbnail

    Feature Generation from Knowledge Graphs

    July 30, 2015  |  Matt Gardner
    A lot of attention has recently been given to the creation of large knowledge bases that contain millions of facts about people, things, and places in the world. In this talk I present methods for using these knowledge bases to generate features for machine learning models. These methods view the knowledge base…
  • Consciousness in Biological and Artificial Brains Thumbnail

    Consciousness in Biological and Artificial Brains

    July 10, 2015  |  Christof Koch
    Human and non-human animals not only act in the world but are capable of conscious experience. That is, it feels like something to have a brain and be cold, angry or see red. I will discuss the scientific progress that has been achieved over the past decades in characterizing the behavioral and the neuronal…
  • Machine Learning with Humans In-the-Loop Thumbnail

    Machine Learning with Humans In-the-Loop

    April 21, 2015  |  Karthik Raman
    In this talk I discuss the challenges of learning from data that results from human behavior. I will present new machine learning models and algorithms that explicitly account for the human decision making process and factors underlying it such as human expertise, skills and needs. The talk will also explore how…
  • Going Beyond Fact-Based Question Answering Thumbnail

    Going Beyond Fact-Based Question Answering

    April 7, 2015  |  Erik T. Mueller
    To solve the AI problem, we need to develop systems that go beyond answering fact-based questions. Watson has been hugely successful at answering fact-based questions, but to solve hard AI tasks like passing science tests and understanding narratives, we need to go beyond simple facts. In this talk, I discuss how…
  • Bring Your Own Model: Model-Agnostic Improvements in NLP Thumbnail

    Bring Your Own Model: Model-Agnostic Improvements in NLP

    April 7, 2015  |  Dani Yogatama
    The majority of NLP research focuses on improving NLP systems by designing better model classes (e.g., non-linear models, latent variable models). In this talk, I will describe a complementary approach based on incorporation of linguistic bias and optimization of text representations that is applicable to several…
  • Learning from Large, Structured Examples Thumbnail

    Learning from Large, Structured Examples

    March 31, 2015  |  
    In many real-world applications of AI and machine learning, such as natural language processing, computer vision and knowledge base construction, data sources possess a natural internal structure, which can be exploited to improve predictive accuracy. Sometimes the structure can be very large, containing many…
  • Distantly Supervised Information Extraction Using Bootstrapped Patterns Thumbnail

    Distantly Supervised Information Extraction Using Bootstrapped Patterns

    March 27, 2015  |  Sonal Gupta
    Although most work in information extraction (IE) focuses on tasks that have abundant training data, in practice, many IE problems do not have any supervised training data. State-of-the-art supervised techniques like conditional random fields are impractical for such real world applications because: (1) they…
  • Exploiting Parallel News Streams for Relation Extraction Thumbnail

    Exploiting Parallel News Streams for Relation Extraction

    March 17, 2015  |  Congle Zhang
    Most approaches to relation extraction, the task of extracting ground facts from natural language text, are based on machine learning and thus starved by scarce training data. Manual annotation is too expensive to scale to a comprehensive set of relations. Distant supervision, which automatically creates training…