diff --git a/gtsam/discrete/DiscreteLookupDAG.cpp b/gtsam/discrete/DiscreteLookupDAG.cpp
index 11900b502..4d02e007b 100644
--- a/gtsam/discrete/DiscreteLookupDAG.cpp
+++ b/gtsam/discrete/DiscreteLookupDAG.cpp
@@ -29,12 +29,90 @@ using std::vector;
 
 namespace gtsam {
 
+/* ************************************************************************** */
+// TODO(dellaert): copy/paste from DiscreteConditional.cpp :-(
+void DiscreteLookupTable::print(const std::string& s,
+                                const KeyFormatter& formatter) const {
+  using std::cout;
+  using std::endl;
+
+  cout << s << " g( ";
+  for (const_iterator it = beginFrontals(); it != endFrontals(); ++it) {
+    cout << formatter(*it) << " ";
+  }
+  if (nrParents()) {
+    cout << "; ";
+    for (const_iterator it = beginParents(); it != endParents(); ++it) {
+      cout << formatter(*it) << " ";
+    }
+  }
+  cout << "):\n";
+  ADT::print("", formatter);
+  cout << endl;
+}
+
+/* ************************************************************************** */
+void DiscreteLookupTable::argmaxInPlace(DiscreteValues* values) const {
+  ADT pFS = choose(*values, true);  // P(F|S=parentsValues)
+
+  // Initialize
+  DiscreteValues mpe;
+  double maxP = 0;
+
+  // Get all Possible Configurations
+  const auto allPosbValues = frontalAssignments();
+
+  // Find the maximum
+  for (const auto& frontalVals : allPosbValues) {
+    double pValueS = pFS(frontalVals);  // P(F=value|S=parentsValues)
+    // Update maximum solution if better
+    if (pValueS > maxP) {
+      maxP = pValueS;
+      mpe = frontalVals;
+    }
+  }
+
+  // set values (inPlace) to maximum
+  for (Key j : frontals()) {
+    (*values)[j] = mpe[j];
+  }
+}
+
+/* ************************************************************************** */
+size_t DiscreteLookupTable::argmax(const DiscreteValues& parentsValues) const {
+  ADT pFS = choose(parentsValues, true);  // P(F|S=parentsValues)
+
+  // Then, find the max over all remaining
+  // TODO(Duy): only works for one key now, seems horribly slow this way
+  size_t mpe = 0;
+  double maxP = 0;
+  DiscreteValues frontals;
+  assert(nrFrontals() == 1);
+  Key j = (firstFrontalKey());
+  for (size_t value = 0; value < cardinality(j); value++) {
+    frontals[j] = value;
+    double pValueS = pFS(frontals);  // P(F=value|S=parentsValues)
+    // Update MPE solution if better
+    if (pValueS > maxP) {
+      maxP = pValueS;
+      mpe = value;
+    }
+  }
+  return mpe;
+}
+
 /* ************************************************************************** */
 DiscreteLookupDAG DiscreteLookupDAG::FromBayesNet(
     const DiscreteBayesNet& bayesNet) {
   DiscreteLookupDAG dag;
   for (auto&& conditional : bayesNet) {
-    dag.push_back(conditional);
+    if (auto lookupTable =
+            std::dynamic_pointer_cast<DiscreteLookupTable>(conditional)) {
+      dag.push_back(lookupTable);
+    } else {
+      throw std::runtime_error(
+          "DiscreteFactorGraph::maxProduct: Expected look up table.");
+    }
   }
   return dag;
 }
diff --git a/gtsam/discrete/DiscreteLookupDAG.h b/gtsam/discrete/DiscreteLookupDAG.h
index 3b0a5770d..f077a13d9 100644
--- a/gtsam/discrete/DiscreteLookupDAG.h
+++ b/gtsam/discrete/DiscreteLookupDAG.h
@@ -37,9 +37,41 @@ class DiscreteBayesNet;
  * Inherits from discrete conditional for convenience, but is not normalized.
  * Is used in the max-product algorithm.
  */
-// Typedef for backwards compatibility
-// TODO(Varun): Remove
-using DiscreteLookupTable = DiscreteConditional;
+class GTSAM_EXPORT DiscreteLookupTable : public DiscreteConditional {
+ public:
+  using This = DiscreteLookupTable;
+  using shared_ptr = std::shared_ptr<This>;
+  using BaseConditional = Conditional<DecisionTreeFactor, This>;
+
+  /**
+   * @brief Construct a new Discrete Lookup Table object
+   *
+   * @param nFrontals number of frontal variables
+   * @param keys a sorted list of gtsam::Keys
+   * @param potentials the algebraic decision tree with lookup values
+   */
+  DiscreteLookupTable(size_t nFrontals, const DiscreteKeys& keys,
+                      const ADT& potentials)
+      : DiscreteConditional(nFrontals, keys, potentials) {}
+
+  /// GTSAM-style print
+  void print(
+      const std::string& s = "Discrete Lookup Table: ",
+      const KeyFormatter& formatter = DefaultKeyFormatter) const override;
+
+  /**
+   * @brief return assignment for single frontal variable that maximizes value.
+   * @param parentsValues Known assignments for the parents.
+   * @return maximizing assignment for the frontal variable.
+   */
+  size_t argmax(const DiscreteValues& parentsValues) const;
+
+  /**
+   * @brief Calculate assignment for frontal variables that maximizes value.
+   * @param (in/out) parentsValues Known assignments for the parents.
+   */
+  void argmaxInPlace(DiscreteValues* parentsValues) const;
+};
 
 /** A DAG made from lookup tables, as defined above. */
 class GTSAM_EXPORT DiscreteLookupDAG : public BayesNet<DiscreteLookupTable> {
diff --git a/gtsam/discrete/discrete.i b/gtsam/discrete/discrete.i
index 0f34840bf..0bdebd0e1 100644
--- a/gtsam/discrete/discrete.i
+++ b/gtsam/discrete/discrete.i
@@ -262,6 +262,15 @@ class DiscreteBayesTree {
 
 #include <gtsam/discrete/DiscreteLookupDAG.h>
 
+class DiscreteLookupTable : gtsam::DiscreteConditional{
+  DiscreteLookupTable(size_t nFrontals, const gtsam::DiscreteKeys& keys,
+                      const gtsam::DecisionTreeFactor::ADT& potentials);
+  void print(string s = "Discrete Lookup Table: ",
+             const gtsam::KeyFormatter& keyFormatter =
+                 gtsam::DefaultKeyFormatter) const;
+  size_t argmax(const gtsam::DiscreteValues& parentsValues) const;
+};
+
 class DiscreteLookupDAG {
   DiscreteLookupDAG();
   void push_back(const gtsam::DiscreteLookupTable* table);