function [marginals, loglik] = enter_soft_evidence(engine, CPDpot, observed, pot_type, filter) % ENTER_SOFT_EVIDENCE Add the specified soft evidence to the network (ff) % [marginals, loglik] = enter_soft_evidence(engine, CPDpot, observed, pot_type, filter) assert(pot_type == 'd'); [ss T] = size(CPDpot); fwd = cell(ss,T); hnodes = engine.hnodes(:)'; onodes = engine.onodes(:)'; bnet = bnet_from_engine(engine); ns = bnet.node_sizes; onodes2 = [onodes onodes+ss]; ns(onodes2) = 1; logscale = zeros(1,T); H = length(hnodes); local_logscale = zeros(1,ss); obschild = zeros(1,ss); for i=hnodes ocs = myintersect(children(bnet.dag, i), onodes); assert(length(ocs)==1); obschild(i) = ocs(1); end t = 1; for i=hnodes fwd{i,t} = CPDpot{i,t}; c = obschild(i); temp = pot_to_marginal(CPDpot{c,t}); ev = dpot(i, ns(i), temp.T); fwd{i,t} = multiply_by_pot(fwd{i,t}, ev); [fwd{i,t}, local_logscale(i)] = normalize_pot(fwd{i,t}); end logscale(t) = sum(local_logscale); for t=2:T for i=hnodes ps = parents(bnet.dag, i+ss); assert(all(ps<=ss)); % in previous slice prior = CPDpot{i,t}; for p=ps(:)' prior = multiply_by_pot(prior, fwd{p,t-1}); end fwd{i,t} = marginalize_pot(prior, i+ss); fwd{i,t} = set_domain_pot(fwd{i,t}, i); c = obschild(i); temp = pot_to_marginal(CPDpot{c,t}); ev = dpot(i, ns(i), temp.T); fwd{i,t} = multiply_by_pot(fwd{i,t}, ev); [fwd{i,t}, local_logscale(i)] = normalize_pot(fwd{i,t}); end logscale(t) = sum(local_logscale); end loglik = sum(logscale); if filter marginals = fwd; return; end back = cell(ss,T); t = T; for i=hnodes back{i,t} = dpot(i, ns(i)); back{i,t} = set_domain_pot(back{i,t}, i+ss); end for t=T-1:-1:1 for i=hnodes pot = CPDpot{i,t+1}; pot = multiply_by_pot(pot, back{i,t+1}); c = obschild(i); temp = pot_to_marginal(CPDpot{c,t+1}); ev = dpot(i, ns(i), temp.T); pot = multiply_by_pot(pot, ev); back{i,t} = marginalize_pot(pot, i); back{i,t} = normalize_pot(back{i,t}); back{i,t} = set_domain_pot(back{i,t}, i+ss); end end % COMBINE for t=1:T for i=hnodes back{i,t} = set_domain_pot(back{i,t}, i); fwd{i,t} = multiply_by_pot(fwd{i,t}, back{i,t}); marginals{i,t} = normalize_pot(fwd{i,t}); %fwdback{i,t} = normalize_pot(multiply_pots(fwd{i,t}, back{i,t})); end end