The lab's research framework seeks to establish how knowledge of human cognitive systems can enhance intelligent systems for practical engineering tasks, and how information environments can be structured to foster seamless collaboration between human operators and machine agents.

Institutional Affiliations and Funding

The CCL is a multidisciplinary hub, structurally integrated into several core programs and institutes at Indiana University, Bloomington. It maintains active research and administrative affiliations with the following entities:

• Department of Psychological and Brain Sciences
• Luddy School of Informatics, Computing, and Engineering
• Cognitive Science Program
• Neuroscience Program
• Indiana University Network Science Institute (IUNI)
• Computational Linguistics Program

Physically, the laboratory is situated on the 6th floor of the Geological Sciences Building within the Center for Computational Cognition and Perception.

The empirical and computational operations conducted by the lab are supported by competitive federal research grants and corporate foundations, including:

• The National Science Foundation (NSF)
• The National Institutes of Health (NIH)
• The Institute of Education Sciences (IES) under the United States Department of Education
• The Indiana Clinical and Translational Sciences Institute (CTSI)
• Google Research

Core Research Operations and Areas of Focus

The laboratory executes high-dimensional computational simulations alongside behavioral experiments to deconstruct semantic memory and cognitive functionality. Its operational research focus spans several key domains:

1. Vector Space Models and Distributional Semantics

The lab investigates how human minds extract meaning from the statistical regularities of linguistic environments. Researchers build and evaluate high-dimensional vector space models (word embeddings) to simulate how lexical items acquire semantic properties through contextual distribution. A key aspect of this work involves auditing the limitations of current error-driven learning architectures in machine learning?such as analyzing the effects of "catastrophic forgetting" when models encounter distinct word senses sequentially rather than in a shuffled distribution.

2. Multi-Modal Information Integration

Recognizing that human cognition does not rely on text alone, the lab develops computational frameworks that synthesize linguistic data with perceptual, sensory, and motor streams. By building models that map how physical sensory inputs interact with linguistic representations, the lab informs the engineering of sensor-based AI systems, such as mobile devices, augmented reality frameworks, and autonomous vehicles that must continuously interpret physical environments alongside language.

3. Exemplar-Based Semantic Memory

Departing from traditional AI prototype abstractions?where a word or concept is stored as a single static average vector?the CCL develops exemplar-based semantic models. These systems preserve episodic traces of individual language encounters, applying nonlinear activation across matching exemplars when prompted by a linguistic cue. This allows the computational models to resolve homonyms and dynamically extract both dominant and subordinate word meanings based on context, mirroring human cognitive flexibility.

4. Continuous Learning and Cognitive Development

Through participation in initiatives like the Emerging Areas of Research Institute titled "Learning: Machines, Brains, & Children," the lab studies how child language acquisition can guide continuous ML learning models. Unlike traditional AI models trained on static, balanced datasets, children learn continuously within dynamic, shifting, and highly unstructured environments. The lab uses large-scale linguistic and video data captured from natural child environments to evaluate how simple distributional models acquire vocabulary and construct concepts over time.

5. Memory Search, Foraging, and Clinical Applications

The lab models the exact algorithmic mechanisms humans employ to navigate and retrieve information from semantic memory during production tasks, such as free association and category fluency drills. The research demonstrates that semantic retrieval functions similarly to optimal foraging algorithms utilized by organisms searching for spatial resources in physical nature. This research is extended to simulate memory illusions, false memories, and cognitive decline, providing formal baselines to understand learning disabilities and developmental disorders.

Facilities and Infrastructure

To execute data-heavy, large-scale computational cognitive modeling alongside empirical human testing, the CCL maintains a specialized facility split between high-performance computing resources and precise behavioral testing instrumentation:

• High-Performance Computing Infrastructure: The lab utilizes specialized local computational clusters customized for intensive simulation workloads. For massive, high-dimensional data mining and multi-modal semantic modeling, the lab leverages Indiana University?s advanced supercomputing grid, allowing the deployment of algorithms that process terabytes of naturalistic linguistic data.
• Behavioral Testing Enclosures: The physical lab space contains 20 individual subject testing booths designed to isolate participants during behavioral, psychological, and language learning experiments.
• Eye-Tracking Instrumentation: The facility is equipped with an SR Research Eyelink 2K remote optics high-speed eye tracker. This system supports monocular and binocular tracking at temporal resolutions up to 2000Hz, capturing data points every half millisecond. This allows researchers to track micro-saccades and visual attention during reading or image comprehension tasks to evaluate real-time cognitive processing.
• Biological Motion Capture: The lab features a specialized Gold-standard biological motion capture system to record and analyze physical movement, supporting projects that map motor controls and physical actions to perceptual and cognitive models.

Practical Applications, Products, and Services

While the Cognitive Computing Lab is a academic research entity rather than a commercial software vendor, it outputs tangible open-source algorithms, methodologies, and datasets that function as computational products for the broader scientific and engineering community. These include:

• Intelligent Search and Information Retrieval Algorithms: Applying cognitive search optimization profiles (such as semantic foraging models) to design search engines that predict user intent and retrieve information based on contextual human memory structures rather than exact keyword or naive embedding matches.
• Knowledge Abstraction Systems: Computational tools capable of scanning unorganized, large-scale text and image databases to extract structured knowledge representations, map relationships, and perform automated data mining.
• Cognitive Diagnostics and Educational Regimens: Mathematical formulations of conceptual learning that establish baseline metrics for typical language development. These baselines serve as analytical models to identify developmental disorders and optimize training regimens for individuals with learning disabilities.
• AI-Based Learning Technologies: Frameworks designed to optimally structure data and interactive pipelines within educational software, matching the statistical ingestion rates of human cognitive systems to maximize retention and skill acquisition.