It is 1994, and the cocktail-party problem has been formalized to its skeleton: two people speak at once, two microphones each record an unknown mixture, and the task is to recover both voices with no model of either speaker and no knowledge of the mixing — blind separation. The key is an assumption of almost impudent weakness: the sources are statistically independent, and independence is all you use. Gaussian statistics provably cannot do it — second-order correlations leave a rotation undetermined — so the algorithm must consume higher-order structure, the non-Gaussianity of real speech. Derive the learning rule, characterize what remains unidentifiable (order and scale), and separate real recordings. Get it wrong and hearing aids, EEG analysis, and every crowded-room interface stay deaf to the voice that matters.
Chose The graphical model — right call.
Jordan's career-long project unifying probabilistic graphical models gives him a natural theoretical lens for framing independent component analysis as a latent-variable model where the sources are the hidden, mutually independent variables to be recovered — a formalization that clarifies exactly why Gaussian sources are unidentifiable and non-Gaussian ones are not. His broad statistical rigor and comfort translating one inference framework into another would let him derive or at least motivate the higher-order learning rule this problem requires. He was not among the researchers directly developing ICA algorithms in the early-to-mid 1990s, and his own major publications in this period center on different graphical-model and mixture-of-experts architectures, so the concrete blind-separation deliverable belongs to specialists working directly on the problem rather than to Jordan himself.
Mukherjee's work on learning theory, topology, and geometry of data, and Bayesian approaches to statistical learning, offers general mathematical sophistication relevant to reasoning about identifiability and estimation limits in blind source separation, though he never worked on independent component analysis, the cocktail-party problem, or higher-order-statistics-based separation specifically. His topological and geometric instincts could in principle characterize the structure of the solution space this problem's ambiguity describes, but the connection remains conceptual rather than a direct application of any published result of his. The concrete 1994 blind-separation deliverable this problem specifies remains outside his own body of published research. His actual technical contributions lie in a genuinely different, later corner of statistical learning theory than this problem occupies.
Battle #165 · 8/10/2026, 11:41:26 AM · this result is deterministic: the same two personas on this problem always resolve the same way.